| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 5.0, |
| "eval_steps": 210, |
| "global_step": 1050, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0, |
| "eval_loss": 2.608412027359009, |
| "eval_ppl": 13.57747, |
| "eval_runtime": 11.2346, |
| "eval_samples_per_second": 7.922, |
| "eval_steps_per_second": 2.047, |
| "memory/device_reserved (GiB)": 36.59, |
| "memory/max_active (GiB)": 34.31, |
| "memory/max_allocated (GiB)": 34.31, |
| "step": 0 |
| }, |
| { |
| "epoch": 0.004761904761904762, |
| "grad_norm": 2.359196662902832, |
| "learning_rate": 0.0, |
| "loss": 2.8488729000091553, |
| "memory/device_reserved (GiB)": 38.92, |
| "memory/max_active (GiB)": 38.08, |
| "memory/max_allocated (GiB)": 38.08, |
| "ppl": 17.26831, |
| "step": 1, |
| "tokens/total": 4096, |
| "tokens/train_per_sec_per_gpu": 6.14, |
| "tokens/trainable": 251 |
| }, |
| { |
| "epoch": 0.009523809523809525, |
| "grad_norm": 1.6851783990859985, |
| "learning_rate": 2.0000000000000003e-06, |
| "loss": 2.6969332695007324, |
| "memory/device_reserved (GiB)": 39.14, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 14.83417, |
| "step": 2, |
| "tokens/total": 8192, |
| "tokens/train_per_sec_per_gpu": 18.92, |
| "tokens/trainable": 606 |
| }, |
| { |
| "epoch": 0.014285714285714285, |
| "grad_norm": 2.1991381645202637, |
| "learning_rate": 4.000000000000001e-06, |
| "loss": 2.4316632747650146, |
| "memory/device_reserved (GiB)": 39.16, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 11.37779, |
| "step": 3, |
| "tokens/total": 12288, |
| "tokens/train_per_sec_per_gpu": 14.49, |
| "tokens/trainable": 869 |
| }, |
| { |
| "epoch": 0.01904761904761905, |
| "grad_norm": 1.6060373783111572, |
| "learning_rate": 6e-06, |
| "loss": 2.2152516841888428, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 9.16372, |
| "step": 4, |
| "tokens/total": 16384, |
| "tokens/train_per_sec_per_gpu": 17.79, |
| "tokens/trainable": 1191 |
| }, |
| { |
| "epoch": 0.023809523809523808, |
| "grad_norm": 3.0328783988952637, |
| "learning_rate": 8.000000000000001e-06, |
| "loss": 2.76442289352417, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 15.86988, |
| "step": 5, |
| "tokens/total": 20480, |
| "tokens/train_per_sec_per_gpu": 9.83, |
| "tokens/trainable": 1449 |
| }, |
| { |
| "epoch": 0.02857142857142857, |
| "grad_norm": 1.5438529253005981, |
| "learning_rate": 1e-05, |
| "loss": 2.1510019302368164, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 8.59346, |
| "step": 6, |
| "tokens/total": 24576, |
| "tokens/train_per_sec_per_gpu": 15.98, |
| "tokens/trainable": 1808 |
| }, |
| { |
| "epoch": 0.03333333333333333, |
| "grad_norm": 1.570513367652893, |
| "learning_rate": 9.99043062200957e-06, |
| "loss": 2.0862762928009033, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 8.05487, |
| "step": 7, |
| "tokens/total": 28672, |
| "tokens/train_per_sec_per_gpu": 16.35, |
| "tokens/trainable": 2100 |
| }, |
| { |
| "epoch": 0.0380952380952381, |
| "grad_norm": 1.9097869396209717, |
| "learning_rate": 9.98086124401914e-06, |
| "loss": 2.539433002471924, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 12.67248, |
| "step": 8, |
| "tokens/total": 32768, |
| "tokens/train_per_sec_per_gpu": 15.58, |
| "tokens/trainable": 2405 |
| }, |
| { |
| "epoch": 0.04285714285714286, |
| "grad_norm": 2.1091222763061523, |
| "learning_rate": 9.971291866028709e-06, |
| "loss": 2.706951856613159, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 14.98353, |
| "step": 9, |
| "tokens/total": 36864, |
| "tokens/train_per_sec_per_gpu": 19.27, |
| "tokens/trainable": 2724 |
| }, |
| { |
| "epoch": 0.047619047619047616, |
| "grad_norm": 1.995006799697876, |
| "learning_rate": 9.961722488038278e-06, |
| "loss": 2.5782675743103027, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 13.17429, |
| "step": 10, |
| "tokens/total": 40960, |
| "tokens/train_per_sec_per_gpu": 16.29, |
| "tokens/trainable": 3076 |
| }, |
| { |
| "epoch": 0.05238095238095238, |
| "grad_norm": 1.8370505571365356, |
| "learning_rate": 9.952153110047848e-06, |
| "loss": 2.5256881713867188, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 12.49949, |
| "step": 11, |
| "tokens/total": 45056, |
| "tokens/train_per_sec_per_gpu": 13.35, |
| "tokens/trainable": 3360 |
| }, |
| { |
| "epoch": 0.05714285714285714, |
| "grad_norm": 1.9744315147399902, |
| "learning_rate": 9.942583732057417e-06, |
| "loss": 3.0185282230377197, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 20.46116, |
| "step": 12, |
| "tokens/total": 49152, |
| "tokens/train_per_sec_per_gpu": 12.19, |
| "tokens/trainable": 3645 |
| }, |
| { |
| "epoch": 0.06190476190476191, |
| "grad_norm": 21.56676483154297, |
| "learning_rate": 9.933014354066987e-06, |
| "loss": 2.959587574005127, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 19.29001, |
| "step": 13, |
| "tokens/total": 53248, |
| "tokens/train_per_sec_per_gpu": 8.84, |
| "tokens/trainable": 3879 |
| }, |
| { |
| "epoch": 0.06666666666666667, |
| "grad_norm": 2.027482509613037, |
| "learning_rate": 9.923444976076556e-06, |
| "loss": 2.170783042907715, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 8.76514, |
| "step": 14, |
| "tokens/total": 57344, |
| "tokens/train_per_sec_per_gpu": 14.41, |
| "tokens/trainable": 4225 |
| }, |
| { |
| "epoch": 0.07142857142857142, |
| "grad_norm": 2.194688558578491, |
| "learning_rate": 9.913875598086125e-06, |
| "loss": 2.2357940673828125, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 9.35391, |
| "step": 15, |
| "tokens/total": 61440, |
| "tokens/train_per_sec_per_gpu": 13.21, |
| "tokens/trainable": 4571 |
| }, |
| { |
| "epoch": 0.0761904761904762, |
| "grad_norm": 1.9835619926452637, |
| "learning_rate": 9.904306220095695e-06, |
| "loss": 2.5124027729034424, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 12.33453, |
| "step": 16, |
| "tokens/total": 65536, |
| "tokens/train_per_sec_per_gpu": 14.99, |
| "tokens/trainable": 4902 |
| }, |
| { |
| "epoch": 0.08095238095238096, |
| "grad_norm": 2.9243738651275635, |
| "learning_rate": 9.894736842105264e-06, |
| "loss": 3.1188597679138184, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 22.62057, |
| "step": 17, |
| "tokens/total": 69632, |
| "tokens/train_per_sec_per_gpu": 9.42, |
| "tokens/trainable": 5144 |
| }, |
| { |
| "epoch": 0.08571428571428572, |
| "grad_norm": 1.6045500040054321, |
| "learning_rate": 9.885167464114833e-06, |
| "loss": 2.362255334854126, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 10.61486, |
| "step": 18, |
| "tokens/total": 73728, |
| "tokens/train_per_sec_per_gpu": 15.12, |
| "tokens/trainable": 5519 |
| }, |
| { |
| "epoch": 0.09047619047619047, |
| "grad_norm": 2.464662551879883, |
| "learning_rate": 9.875598086124403e-06, |
| "loss": 2.4410171508789062, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 11.48472, |
| "step": 19, |
| "tokens/total": 77824, |
| "tokens/train_per_sec_per_gpu": 23.74, |
| "tokens/trainable": 5856 |
| }, |
| { |
| "epoch": 0.09523809523809523, |
| "grad_norm": 2.090559482574463, |
| "learning_rate": 9.866028708133972e-06, |
| "loss": 2.2410788536071777, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 9.40347, |
| "step": 20, |
| "tokens/total": 81920, |
| "tokens/train_per_sec_per_gpu": 21.14, |
| "tokens/trainable": 6239 |
| }, |
| { |
| "epoch": 0.1, |
| "grad_norm": 1.6499180793762207, |
| "learning_rate": 9.856459330143542e-06, |
| "loss": 1.750786304473877, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 5.75913, |
| "step": 21, |
| "tokens/total": 86016, |
| "tokens/train_per_sec_per_gpu": 12.88, |
| "tokens/trainable": 6562 |
| }, |
| { |
| "epoch": 0.10476190476190476, |
| "grad_norm": 2.06025767326355, |
| "learning_rate": 9.846889952153111e-06, |
| "loss": 2.453636646270752, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 11.63057, |
| "step": 22, |
| "tokens/total": 90112, |
| "tokens/train_per_sec_per_gpu": 14.53, |
| "tokens/trainable": 6881 |
| }, |
| { |
| "epoch": 0.10952380952380952, |
| "grad_norm": 2.1834657192230225, |
| "learning_rate": 9.83732057416268e-06, |
| "loss": 2.3299005031585693, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 10.27692, |
| "step": 23, |
| "tokens/total": 94208, |
| "tokens/train_per_sec_per_gpu": 16.89, |
| "tokens/trainable": 7169 |
| }, |
| { |
| "epoch": 0.11428571428571428, |
| "grad_norm": 2.7063045501708984, |
| "learning_rate": 9.82775119617225e-06, |
| "loss": 2.135418653488159, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 8.46059, |
| "step": 24, |
| "tokens/total": 98304, |
| "tokens/train_per_sec_per_gpu": 15.24, |
| "tokens/trainable": 7424 |
| }, |
| { |
| "epoch": 0.11904761904761904, |
| "grad_norm": 2.213444232940674, |
| "learning_rate": 9.81818181818182e-06, |
| "loss": 2.230344295501709, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 9.30307, |
| "step": 25, |
| "tokens/total": 102400, |
| "tokens/train_per_sec_per_gpu": 12.01, |
| "tokens/trainable": 7700 |
| }, |
| { |
| "epoch": 0.12380952380952381, |
| "grad_norm": 2.1053807735443115, |
| "learning_rate": 9.808612440191389e-06, |
| "loss": 1.8845914602279663, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 6.58366, |
| "step": 26, |
| "tokens/total": 106496, |
| "tokens/train_per_sec_per_gpu": 12.01, |
| "tokens/trainable": 7969 |
| }, |
| { |
| "epoch": 0.12857142857142856, |
| "grad_norm": 2.621504783630371, |
| "learning_rate": 9.799043062200958e-06, |
| "loss": 2.465123176574707, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 11.76493, |
| "step": 27, |
| "tokens/total": 110592, |
| "tokens/train_per_sec_per_gpu": 12.26, |
| "tokens/trainable": 8221 |
| }, |
| { |
| "epoch": 0.13333333333333333, |
| "grad_norm": 2.0597894191741943, |
| "learning_rate": 9.789473684210527e-06, |
| "loss": 2.136955738067627, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 8.4736, |
| "step": 28, |
| "tokens/total": 114688, |
| "tokens/train_per_sec_per_gpu": 17.31, |
| "tokens/trainable": 8540 |
| }, |
| { |
| "epoch": 0.1380952380952381, |
| "grad_norm": 2.4883816242218018, |
| "learning_rate": 9.779904306220097e-06, |
| "loss": 1.710213541984558, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 5.53014, |
| "step": 29, |
| "tokens/total": 118784, |
| "tokens/train_per_sec_per_gpu": 16.86, |
| "tokens/trainable": 8852 |
| }, |
| { |
| "epoch": 0.14285714285714285, |
| "grad_norm": 3.3769397735595703, |
| "learning_rate": 9.770334928229666e-06, |
| "loss": 2.047621011734009, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 7.74944, |
| "step": 30, |
| "tokens/total": 122880, |
| "tokens/train_per_sec_per_gpu": 10.15, |
| "tokens/trainable": 9091 |
| }, |
| { |
| "epoch": 0.14761904761904762, |
| "grad_norm": 2.247663736343384, |
| "learning_rate": 9.760765550239236e-06, |
| "loss": 1.5432257652282715, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 4.67966, |
| "step": 31, |
| "tokens/total": 126976, |
| "tokens/train_per_sec_per_gpu": 9.37, |
| "tokens/trainable": 9355 |
| }, |
| { |
| "epoch": 0.1523809523809524, |
| "grad_norm": 2.2224535942077637, |
| "learning_rate": 9.751196172248805e-06, |
| "loss": 2.1464388370513916, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 8.55434, |
| "step": 32, |
| "tokens/total": 131072, |
| "tokens/train_per_sec_per_gpu": 12.68, |
| "tokens/trainable": 9634 |
| }, |
| { |
| "epoch": 0.15714285714285714, |
| "grad_norm": 3.218093156814575, |
| "learning_rate": 9.741626794258374e-06, |
| "loss": 2.0251033306121826, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 7.57689, |
| "step": 33, |
| "tokens/total": 135168, |
| "tokens/train_per_sec_per_gpu": 10.22, |
| "tokens/trainable": 9926 |
| }, |
| { |
| "epoch": 0.1619047619047619, |
| "grad_norm": 3.240586280822754, |
| "learning_rate": 9.732057416267944e-06, |
| "loss": 1.7320709228515625, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 5.65235, |
| "step": 34, |
| "tokens/total": 139264, |
| "tokens/train_per_sec_per_gpu": 15.82, |
| "tokens/trainable": 10196 |
| }, |
| { |
| "epoch": 0.16666666666666666, |
| "grad_norm": 2.352398633956909, |
| "learning_rate": 9.722488038277513e-06, |
| "loss": 1.7620348930358887, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 5.82428, |
| "step": 35, |
| "tokens/total": 143360, |
| "tokens/train_per_sec_per_gpu": 18.11, |
| "tokens/trainable": 10539 |
| }, |
| { |
| "epoch": 0.17142857142857143, |
| "grad_norm": 2.2571511268615723, |
| "learning_rate": 9.712918660287082e-06, |
| "loss": 1.7542566061019897, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 5.77915, |
| "step": 36, |
| "tokens/total": 147456, |
| "tokens/train_per_sec_per_gpu": 17.97, |
| "tokens/trainable": 10838 |
| }, |
| { |
| "epoch": 0.1761904761904762, |
| "grad_norm": 2.1978981494903564, |
| "learning_rate": 9.703349282296652e-06, |
| "loss": 1.33998441696167, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 3.81898, |
| "step": 37, |
| "tokens/total": 151552, |
| "tokens/train_per_sec_per_gpu": 14.33, |
| "tokens/trainable": 11211 |
| }, |
| { |
| "epoch": 0.18095238095238095, |
| "grad_norm": 2.0772902965545654, |
| "learning_rate": 9.693779904306221e-06, |
| "loss": 1.467942237854004, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 4.34029, |
| "step": 38, |
| "tokens/total": 155648, |
| "tokens/train_per_sec_per_gpu": 12.64, |
| "tokens/trainable": 11467 |
| }, |
| { |
| "epoch": 0.18571428571428572, |
| "grad_norm": 2.1938557624816895, |
| "learning_rate": 9.68421052631579e-06, |
| "loss": 1.2810404300689697, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 3.60038, |
| "step": 39, |
| "tokens/total": 159744, |
| "tokens/train_per_sec_per_gpu": 17.24, |
| "tokens/trainable": 11792 |
| }, |
| { |
| "epoch": 0.19047619047619047, |
| "grad_norm": 1.9610248804092407, |
| "learning_rate": 9.67464114832536e-06, |
| "loss": 1.304593801498413, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 3.68619, |
| "step": 40, |
| "tokens/total": 163840, |
| "tokens/train_per_sec_per_gpu": 14.73, |
| "tokens/trainable": 12086 |
| }, |
| { |
| "epoch": 0.19523809523809524, |
| "grad_norm": 2.6710472106933594, |
| "learning_rate": 9.66507177033493e-06, |
| "loss": 1.5237510204315186, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 4.58941, |
| "step": 41, |
| "tokens/total": 167936, |
| "tokens/train_per_sec_per_gpu": 12.31, |
| "tokens/trainable": 12326 |
| }, |
| { |
| "epoch": 0.2, |
| "grad_norm": 2.024487257003784, |
| "learning_rate": 9.655502392344497e-06, |
| "loss": 1.2347090244293213, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 3.43738, |
| "step": 42, |
| "tokens/total": 172032, |
| "tokens/train_per_sec_per_gpu": 14.25, |
| "tokens/trainable": 12720 |
| }, |
| { |
| "epoch": 0.20476190476190476, |
| "grad_norm": 2.842580556869507, |
| "learning_rate": 9.645933014354068e-06, |
| "loss": 1.609395980834961, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 4.99979, |
| "step": 43, |
| "tokens/total": 176128, |
| "tokens/train_per_sec_per_gpu": 10.82, |
| "tokens/trainable": 12928 |
| }, |
| { |
| "epoch": 0.20952380952380953, |
| "grad_norm": 2.1989729404449463, |
| "learning_rate": 9.636363636363638e-06, |
| "loss": 1.1074391603469849, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 3.0266, |
| "step": 44, |
| "tokens/total": 180224, |
| "tokens/train_per_sec_per_gpu": 11.88, |
| "tokens/trainable": 13185 |
| }, |
| { |
| "epoch": 0.21428571428571427, |
| "grad_norm": 5.464274883270264, |
| "learning_rate": 9.626794258373207e-06, |
| "loss": 0.9748563170433044, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 2.65079, |
| "step": 45, |
| "tokens/total": 184320, |
| "tokens/train_per_sec_per_gpu": 12.97, |
| "tokens/trainable": 13515 |
| }, |
| { |
| "epoch": 0.21904761904761905, |
| "grad_norm": 2.9592573642730713, |
| "learning_rate": 9.617224880382776e-06, |
| "loss": 1.2125438451766968, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 3.36203, |
| "step": 46, |
| "tokens/total": 188416, |
| "tokens/train_per_sec_per_gpu": 10.9, |
| "tokens/trainable": 13753 |
| }, |
| { |
| "epoch": 0.22380952380952382, |
| "grad_norm": 2.570880174636841, |
| "learning_rate": 9.607655502392346e-06, |
| "loss": 1.0907056331634521, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 2.97637, |
| "step": 47, |
| "tokens/total": 192512, |
| "tokens/train_per_sec_per_gpu": 12.15, |
| "tokens/trainable": 14051 |
| }, |
| { |
| "epoch": 0.22857142857142856, |
| "grad_norm": 2.5704782009124756, |
| "learning_rate": 9.598086124401913e-06, |
| "loss": 0.9632624387741089, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 2.62023, |
| "step": 48, |
| "tokens/total": 196608, |
| "tokens/train_per_sec_per_gpu": 7.27, |
| "tokens/trainable": 14273 |
| }, |
| { |
| "epoch": 0.23333333333333334, |
| "grad_norm": 2.3074774742126465, |
| "learning_rate": 9.588516746411483e-06, |
| "loss": 0.811287522315979, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 2.2508, |
| "step": 49, |
| "tokens/total": 200704, |
| "tokens/train_per_sec_per_gpu": 21.78, |
| "tokens/trainable": 14641 |
| }, |
| { |
| "epoch": 0.23809523809523808, |
| "grad_norm": 1.8212100267410278, |
| "learning_rate": 9.578947368421054e-06, |
| "loss": 0.8015825748443604, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 2.22907, |
| "step": 50, |
| "tokens/total": 204800, |
| "tokens/train_per_sec_per_gpu": 12.54, |
| "tokens/trainable": 14991 |
| }, |
| { |
| "epoch": 0.24285714285714285, |
| "grad_norm": 2.3841068744659424, |
| "learning_rate": 9.569377990430623e-06, |
| "loss": 0.896703839302063, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 2.45151, |
| "step": 51, |
| "tokens/total": 208896, |
| "tokens/train_per_sec_per_gpu": 15.32, |
| "tokens/trainable": 15302 |
| }, |
| { |
| "epoch": 0.24761904761904763, |
| "grad_norm": 2.0620293617248535, |
| "learning_rate": 9.559808612440193e-06, |
| "loss": 0.7413748502731323, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 2.09882, |
| "step": 52, |
| "tokens/total": 212992, |
| "tokens/train_per_sec_per_gpu": 13.33, |
| "tokens/trainable": 15566 |
| }, |
| { |
| "epoch": 0.2523809523809524, |
| "grad_norm": 1.9990394115447998, |
| "learning_rate": 9.550239234449762e-06, |
| "loss": 0.6690434217453003, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.95237, |
| "step": 53, |
| "tokens/total": 217088, |
| "tokens/train_per_sec_per_gpu": 9.25, |
| "tokens/trainable": 15884 |
| }, |
| { |
| "epoch": 0.2571428571428571, |
| "grad_norm": 2.1852803230285645, |
| "learning_rate": 9.540669856459331e-06, |
| "loss": 0.7765368819236755, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 2.17393, |
| "step": 54, |
| "tokens/total": 221184, |
| "tokens/train_per_sec_per_gpu": 12.36, |
| "tokens/trainable": 16181 |
| }, |
| { |
| "epoch": 0.2619047619047619, |
| "grad_norm": 2.215074300765991, |
| "learning_rate": 9.5311004784689e-06, |
| "loss": 0.6537740230560303, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.92278, |
| "step": 55, |
| "tokens/total": 225280, |
| "tokens/train_per_sec_per_gpu": 10.6, |
| "tokens/trainable": 16458 |
| }, |
| { |
| "epoch": 0.26666666666666666, |
| "grad_norm": 1.7708910703659058, |
| "learning_rate": 9.52153110047847e-06, |
| "loss": 0.7781250476837158, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 2.17739, |
| "step": 56, |
| "tokens/total": 229376, |
| "tokens/train_per_sec_per_gpu": 12.4, |
| "tokens/trainable": 16742 |
| }, |
| { |
| "epoch": 0.2714285714285714, |
| "grad_norm": 2.0387091636657715, |
| "learning_rate": 9.51196172248804e-06, |
| "loss": 0.5093286633491516, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.66417, |
| "step": 57, |
| "tokens/total": 233472, |
| "tokens/train_per_sec_per_gpu": 9.72, |
| "tokens/trainable": 16957 |
| }, |
| { |
| "epoch": 0.2761904761904762, |
| "grad_norm": 1.7607407569885254, |
| "learning_rate": 9.502392344497609e-06, |
| "loss": 0.6932368278503418, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 2.00018, |
| "step": 58, |
| "tokens/total": 237568, |
| "tokens/train_per_sec_per_gpu": 12.7, |
| "tokens/trainable": 17245 |
| }, |
| { |
| "epoch": 0.28095238095238095, |
| "grad_norm": 2.5780985355377197, |
| "learning_rate": 9.492822966507178e-06, |
| "loss": 0.4341663718223572, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.54368, |
| "step": 59, |
| "tokens/total": 241664, |
| "tokens/train_per_sec_per_gpu": 12.24, |
| "tokens/trainable": 17586 |
| }, |
| { |
| "epoch": 0.2857142857142857, |
| "grad_norm": 2.317253351211548, |
| "learning_rate": 9.483253588516748e-06, |
| "loss": 0.5591937303543091, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.74926, |
| "step": 60, |
| "tokens/total": 245760, |
| "tokens/train_per_sec_per_gpu": 11.24, |
| "tokens/trainable": 17828 |
| }, |
| { |
| "epoch": 0.2904761904761905, |
| "grad_norm": 1.5913702249526978, |
| "learning_rate": 9.473684210526315e-06, |
| "loss": 0.5771949291229248, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.78104, |
| "step": 61, |
| "tokens/total": 249856, |
| "tokens/train_per_sec_per_gpu": 17.34, |
| "tokens/trainable": 18107 |
| }, |
| { |
| "epoch": 0.29523809523809524, |
| "grad_norm": 1.7437083721160889, |
| "learning_rate": 9.464114832535885e-06, |
| "loss": 0.3899308443069458, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.47688, |
| "step": 62, |
| "tokens/total": 253952, |
| "tokens/train_per_sec_per_gpu": 7.46, |
| "tokens/trainable": 18304 |
| }, |
| { |
| "epoch": 0.3, |
| "grad_norm": 1.510191798210144, |
| "learning_rate": 9.454545454545456e-06, |
| "loss": 0.4528539776802063, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.57279, |
| "step": 63, |
| "tokens/total": 258048, |
| "tokens/train_per_sec_per_gpu": 8.28, |
| "tokens/trainable": 18531 |
| }, |
| { |
| "epoch": 0.3047619047619048, |
| "grad_norm": 1.3171509504318237, |
| "learning_rate": 9.444976076555025e-06, |
| "loss": 0.44899454712867737, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.56674, |
| "step": 64, |
| "tokens/total": 262144, |
| "tokens/train_per_sec_per_gpu": 14.26, |
| "tokens/trainable": 18919 |
| }, |
| { |
| "epoch": 0.30952380952380953, |
| "grad_norm": 0.8474973440170288, |
| "learning_rate": 9.435406698564595e-06, |
| "loss": 0.22856196761131287, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.25679, |
| "step": 65, |
| "tokens/total": 266240, |
| "tokens/train_per_sec_per_gpu": 10.41, |
| "tokens/trainable": 19181 |
| }, |
| { |
| "epoch": 0.3142857142857143, |
| "grad_norm": 1.0009809732437134, |
| "learning_rate": 9.425837320574164e-06, |
| "loss": 0.3449355959892273, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.4119, |
| "step": 66, |
| "tokens/total": 270336, |
| "tokens/train_per_sec_per_gpu": 18.8, |
| "tokens/trainable": 19493 |
| }, |
| { |
| "epoch": 0.319047619047619, |
| "grad_norm": 1.1229182481765747, |
| "learning_rate": 9.416267942583734e-06, |
| "loss": 0.3861968219280243, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.47137, |
| "step": 67, |
| "tokens/total": 274432, |
| "tokens/train_per_sec_per_gpu": 10.13, |
| "tokens/trainable": 19742 |
| }, |
| { |
| "epoch": 0.3238095238095238, |
| "grad_norm": 2.035735845565796, |
| "learning_rate": 9.406698564593301e-06, |
| "loss": 0.46546095609664917, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.59275, |
| "step": 68, |
| "tokens/total": 278528, |
| "tokens/train_per_sec_per_gpu": 14.0, |
| "tokens/trainable": 20081 |
| }, |
| { |
| "epoch": 0.32857142857142857, |
| "grad_norm": 0.943077802658081, |
| "learning_rate": 9.39712918660287e-06, |
| "loss": 0.2488720715045929, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.28258, |
| "step": 69, |
| "tokens/total": 282624, |
| "tokens/train_per_sec_per_gpu": 12.21, |
| "tokens/trainable": 20364 |
| }, |
| { |
| "epoch": 0.3333333333333333, |
| "grad_norm": 0.8868593573570251, |
| "learning_rate": 9.387559808612442e-06, |
| "loss": 0.4037240743637085, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.49739, |
| "step": 70, |
| "tokens/total": 286720, |
| "tokens/train_per_sec_per_gpu": 10.54, |
| "tokens/trainable": 20641 |
| }, |
| { |
| "epoch": 0.3380952380952381, |
| "grad_norm": 1.323467493057251, |
| "learning_rate": 9.377990430622011e-06, |
| "loss": 0.32316213846206665, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.38149, |
| "step": 71, |
| "tokens/total": 290816, |
| "tokens/train_per_sec_per_gpu": 15.32, |
| "tokens/trainable": 20995 |
| }, |
| { |
| "epoch": 0.34285714285714286, |
| "grad_norm": 0.9882053136825562, |
| "learning_rate": 9.36842105263158e-06, |
| "loss": 0.24642500281333923, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.27944, |
| "step": 72, |
| "tokens/total": 294912, |
| "tokens/train_per_sec_per_gpu": 12.04, |
| "tokens/trainable": 21353 |
| }, |
| { |
| "epoch": 0.3476190476190476, |
| "grad_norm": 0.7984920144081116, |
| "learning_rate": 9.35885167464115e-06, |
| "loss": 0.25220850110054016, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.28686, |
| "step": 73, |
| "tokens/total": 299008, |
| "tokens/train_per_sec_per_gpu": 19.03, |
| "tokens/trainable": 21721 |
| }, |
| { |
| "epoch": 0.3523809523809524, |
| "grad_norm": 0.8392660617828369, |
| "learning_rate": 9.349282296650718e-06, |
| "loss": 0.2952135503292084, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.34341, |
| "step": 74, |
| "tokens/total": 303104, |
| "tokens/train_per_sec_per_gpu": 10.37, |
| "tokens/trainable": 22012 |
| }, |
| { |
| "epoch": 0.35714285714285715, |
| "grad_norm": 1.0957508087158203, |
| "learning_rate": 9.339712918660287e-06, |
| "loss": 0.3907988965511322, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.47816, |
| "step": 75, |
| "tokens/total": 307200, |
| "tokens/train_per_sec_per_gpu": 11.42, |
| "tokens/trainable": 22253 |
| }, |
| { |
| "epoch": 0.3619047619047619, |
| "grad_norm": 1.0357823371887207, |
| "learning_rate": 9.330143540669856e-06, |
| "loss": 0.25602155923843384, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.29178, |
| "step": 76, |
| "tokens/total": 311296, |
| "tokens/train_per_sec_per_gpu": 11.79, |
| "tokens/trainable": 22517 |
| }, |
| { |
| "epoch": 0.36666666666666664, |
| "grad_norm": 1.4284818172454834, |
| "learning_rate": 9.320574162679427e-06, |
| "loss": 0.3153234124183655, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.3707, |
| "step": 77, |
| "tokens/total": 315392, |
| "tokens/train_per_sec_per_gpu": 8.11, |
| "tokens/trainable": 22760 |
| }, |
| { |
| "epoch": 0.37142857142857144, |
| "grad_norm": 0.8853760361671448, |
| "learning_rate": 9.311004784688997e-06, |
| "loss": 0.2890888452529907, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.33521, |
| "step": 78, |
| "tokens/total": 319488, |
| "tokens/train_per_sec_per_gpu": 9.88, |
| "tokens/trainable": 23027 |
| }, |
| { |
| "epoch": 0.3761904761904762, |
| "grad_norm": 1.0980620384216309, |
| "learning_rate": 9.301435406698566e-06, |
| "loss": 0.27998337149620056, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.32311, |
| "step": 79, |
| "tokens/total": 323584, |
| "tokens/train_per_sec_per_gpu": 14.48, |
| "tokens/trainable": 23322 |
| }, |
| { |
| "epoch": 0.38095238095238093, |
| "grad_norm": 1.0161164999008179, |
| "learning_rate": 9.291866028708136e-06, |
| "loss": 0.3861690163612366, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.47133, |
| "step": 80, |
| "tokens/total": 327680, |
| "tokens/train_per_sec_per_gpu": 15.36, |
| "tokens/trainable": 23639 |
| }, |
| { |
| "epoch": 0.38571428571428573, |
| "grad_norm": 0.9894334077835083, |
| "learning_rate": 9.282296650717703e-06, |
| "loss": 0.2312101125717163, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.26012, |
| "step": 81, |
| "tokens/total": 331776, |
| "tokens/train_per_sec_per_gpu": 10.84, |
| "tokens/trainable": 23899 |
| }, |
| { |
| "epoch": 0.3904761904761905, |
| "grad_norm": 0.8046990633010864, |
| "learning_rate": 9.272727272727273e-06, |
| "loss": 0.1597203016281128, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17318, |
| "step": 82, |
| "tokens/total": 335872, |
| "tokens/train_per_sec_per_gpu": 9.12, |
| "tokens/trainable": 24160 |
| }, |
| { |
| "epoch": 0.3952380952380952, |
| "grad_norm": 0.8290164470672607, |
| "learning_rate": 9.263157894736842e-06, |
| "loss": 0.21786154806613922, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.24341, |
| "step": 83, |
| "tokens/total": 339968, |
| "tokens/train_per_sec_per_gpu": 11.81, |
| "tokens/trainable": 24472 |
| }, |
| { |
| "epoch": 0.4, |
| "grad_norm": 0.7340158820152283, |
| "learning_rate": 9.253588516746413e-06, |
| "loss": 0.1657000184059143, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.18022, |
| "step": 84, |
| "tokens/total": 344064, |
| "tokens/train_per_sec_per_gpu": 11.69, |
| "tokens/trainable": 24724 |
| }, |
| { |
| "epoch": 0.40476190476190477, |
| "grad_norm": 0.6350992321968079, |
| "learning_rate": 9.244019138755983e-06, |
| "loss": 0.18384049832820892, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.20182, |
| "step": 85, |
| "tokens/total": 348160, |
| "tokens/train_per_sec_per_gpu": 16.55, |
| "tokens/trainable": 25031 |
| }, |
| { |
| "epoch": 0.4095238095238095, |
| "grad_norm": 0.9644314050674438, |
| "learning_rate": 9.234449760765552e-06, |
| "loss": 0.23870739340782166, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.26961, |
| "step": 86, |
| "tokens/total": 352256, |
| "tokens/train_per_sec_per_gpu": 11.23, |
| "tokens/trainable": 25335 |
| }, |
| { |
| "epoch": 0.4142857142857143, |
| "grad_norm": 0.8144553303718567, |
| "learning_rate": 9.22488038277512e-06, |
| "loss": 0.18887390196323395, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.20789, |
| "step": 87, |
| "tokens/total": 356352, |
| "tokens/train_per_sec_per_gpu": 12.26, |
| "tokens/trainable": 25618 |
| }, |
| { |
| "epoch": 0.41904761904761906, |
| "grad_norm": 0.6886969804763794, |
| "learning_rate": 9.215311004784689e-06, |
| "loss": 0.20714828372001648, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.23016, |
| "step": 88, |
| "tokens/total": 360448, |
| "tokens/train_per_sec_per_gpu": 14.28, |
| "tokens/trainable": 25916 |
| }, |
| { |
| "epoch": 0.4238095238095238, |
| "grad_norm": 0.6070327162742615, |
| "learning_rate": 9.205741626794258e-06, |
| "loss": 0.15140344202518463, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16347, |
| "step": 89, |
| "tokens/total": 364544, |
| "tokens/train_per_sec_per_gpu": 11.41, |
| "tokens/trainable": 26196 |
| }, |
| { |
| "epoch": 0.42857142857142855, |
| "grad_norm": 1.1441844701766968, |
| "learning_rate": 9.196172248803828e-06, |
| "loss": 0.17098617553710938, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.18647, |
| "step": 90, |
| "tokens/total": 368640, |
| "tokens/train_per_sec_per_gpu": 14.67, |
| "tokens/trainable": 26544 |
| }, |
| { |
| "epoch": 0.43333333333333335, |
| "grad_norm": 0.6829278469085693, |
| "learning_rate": 9.186602870813399e-06, |
| "loss": 0.19169265031814575, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.2113, |
| "step": 91, |
| "tokens/total": 372736, |
| "tokens/train_per_sec_per_gpu": 14.1, |
| "tokens/trainable": 26892 |
| }, |
| { |
| "epoch": 0.4380952380952381, |
| "grad_norm": 0.8788889646530151, |
| "learning_rate": 9.177033492822968e-06, |
| "loss": 0.22554373741149902, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.253, |
| "step": 92, |
| "tokens/total": 376832, |
| "tokens/train_per_sec_per_gpu": 10.81, |
| "tokens/trainable": 27152 |
| }, |
| { |
| "epoch": 0.44285714285714284, |
| "grad_norm": 0.9059855341911316, |
| "learning_rate": 9.167464114832538e-06, |
| "loss": 0.20014435052871704, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.22158, |
| "step": 93, |
| "tokens/total": 380928, |
| "tokens/train_per_sec_per_gpu": 15.86, |
| "tokens/trainable": 27586 |
| }, |
| { |
| "epoch": 0.44761904761904764, |
| "grad_norm": 0.8185603618621826, |
| "learning_rate": 9.157894736842105e-06, |
| "loss": 0.18692737817764282, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.20554, |
| "step": 94, |
| "tokens/total": 385024, |
| "tokens/train_per_sec_per_gpu": 12.94, |
| "tokens/trainable": 27867 |
| }, |
| { |
| "epoch": 0.4523809523809524, |
| "grad_norm": 1.0522328615188599, |
| "learning_rate": 9.148325358851675e-06, |
| "loss": 0.30514901876449585, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.35683, |
| "step": 95, |
| "tokens/total": 389120, |
| "tokens/train_per_sec_per_gpu": 15.6, |
| "tokens/trainable": 28154 |
| }, |
| { |
| "epoch": 0.45714285714285713, |
| "grad_norm": 1.0001094341278076, |
| "learning_rate": 9.138755980861244e-06, |
| "loss": 0.31013065576553345, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.3636, |
| "step": 96, |
| "tokens/total": 393216, |
| "tokens/train_per_sec_per_gpu": 20.17, |
| "tokens/trainable": 28497 |
| }, |
| { |
| "epoch": 0.46190476190476193, |
| "grad_norm": 0.6931004524230957, |
| "learning_rate": 9.129186602870813e-06, |
| "loss": 0.22375206649303436, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.25076, |
| "step": 97, |
| "tokens/total": 397312, |
| "tokens/train_per_sec_per_gpu": 13.8, |
| "tokens/trainable": 28826 |
| }, |
| { |
| "epoch": 0.4666666666666667, |
| "grad_norm": 0.9354197978973389, |
| "learning_rate": 9.119617224880385e-06, |
| "loss": 0.2702368199825287, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.31027, |
| "step": 98, |
| "tokens/total": 401408, |
| "tokens/train_per_sec_per_gpu": 12.0, |
| "tokens/trainable": 29085 |
| }, |
| { |
| "epoch": 0.4714285714285714, |
| "grad_norm": 0.656976044178009, |
| "learning_rate": 9.110047846889954e-06, |
| "loss": 0.16342169046401978, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17753, |
| "step": 99, |
| "tokens/total": 405504, |
| "tokens/train_per_sec_per_gpu": 12.56, |
| "tokens/trainable": 29361 |
| }, |
| { |
| "epoch": 0.47619047619047616, |
| "grad_norm": 0.7472529411315918, |
| "learning_rate": 9.100478468899522e-06, |
| "loss": 0.2030039131641388, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.22508, |
| "step": 100, |
| "tokens/total": 409600, |
| "tokens/train_per_sec_per_gpu": 9.78, |
| "tokens/trainable": 29583 |
| }, |
| { |
| "epoch": 0.48095238095238096, |
| "grad_norm": 0.8217137455940247, |
| "learning_rate": 9.090909090909091e-06, |
| "loss": 0.26452142000198364, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.30281, |
| "step": 101, |
| "tokens/total": 413696, |
| "tokens/train_per_sec_per_gpu": 17.67, |
| "tokens/trainable": 29942 |
| }, |
| { |
| "epoch": 0.4857142857142857, |
| "grad_norm": 0.7122595310211182, |
| "learning_rate": 9.08133971291866e-06, |
| "loss": 0.2568754553794861, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.29288, |
| "step": 102, |
| "tokens/total": 417792, |
| "tokens/train_per_sec_per_gpu": 17.56, |
| "tokens/trainable": 30371 |
| }, |
| { |
| "epoch": 0.49047619047619045, |
| "grad_norm": 1.2763384580612183, |
| "learning_rate": 9.07177033492823e-06, |
| "loss": 0.18725821375846863, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.20594, |
| "step": 103, |
| "tokens/total": 421888, |
| "tokens/train_per_sec_per_gpu": 9.5, |
| "tokens/trainable": 30564 |
| }, |
| { |
| "epoch": 0.49523809523809526, |
| "grad_norm": 0.7632704973220825, |
| "learning_rate": 9.0622009569378e-06, |
| "loss": 0.22999542951583862, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.25859, |
| "step": 104, |
| "tokens/total": 425984, |
| "tokens/train_per_sec_per_gpu": 13.28, |
| "tokens/trainable": 30867 |
| }, |
| { |
| "epoch": 0.5, |
| "grad_norm": 0.5632386803627014, |
| "learning_rate": 9.05263157894737e-06, |
| "loss": 0.15592020750045776, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16873, |
| "step": 105, |
| "tokens/total": 430080, |
| "tokens/train_per_sec_per_gpu": 15.86, |
| "tokens/trainable": 31168 |
| }, |
| { |
| "epoch": 0.5047619047619047, |
| "grad_norm": 0.6003026962280273, |
| "learning_rate": 9.043062200956938e-06, |
| "loss": 0.20485958456993103, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.22735, |
| "step": 106, |
| "tokens/total": 434176, |
| "tokens/train_per_sec_per_gpu": 14.32, |
| "tokens/trainable": 31478 |
| }, |
| { |
| "epoch": 0.5095238095238095, |
| "grad_norm": 0.8614497184753418, |
| "learning_rate": 9.033492822966507e-06, |
| "loss": 0.14312103390693665, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15387, |
| "step": 107, |
| "tokens/total": 438272, |
| "tokens/train_per_sec_per_gpu": 7.5, |
| "tokens/trainable": 31666 |
| }, |
| { |
| "epoch": 0.5142857142857142, |
| "grad_norm": 0.5496851205825806, |
| "learning_rate": 9.023923444976077e-06, |
| "loss": 0.21571555733680725, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.24075, |
| "step": 108, |
| "tokens/total": 442368, |
| "tokens/train_per_sec_per_gpu": 9.59, |
| "tokens/trainable": 32019 |
| }, |
| { |
| "epoch": 0.5190476190476191, |
| "grad_norm": 0.9730420112609863, |
| "learning_rate": 9.014354066985646e-06, |
| "loss": 0.15710116922855377, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17011, |
| "step": 109, |
| "tokens/total": 446464, |
| "tokens/train_per_sec_per_gpu": 11.14, |
| "tokens/trainable": 32262 |
| }, |
| { |
| "epoch": 0.5238095238095238, |
| "grad_norm": 0.7044121026992798, |
| "learning_rate": 9.004784688995216e-06, |
| "loss": 0.15290921926498413, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.16522, |
| "step": 110, |
| "tokens/total": 450560, |
| "tokens/train_per_sec_per_gpu": 12.39, |
| "tokens/trainable": 32499 |
| }, |
| { |
| "epoch": 0.5285714285714286, |
| "grad_norm": 0.8730543851852417, |
| "learning_rate": 8.995215311004785e-06, |
| "loss": 0.2936154305934906, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.34127, |
| "step": 111, |
| "tokens/total": 454656, |
| "tokens/train_per_sec_per_gpu": 10.61, |
| "tokens/trainable": 32752 |
| }, |
| { |
| "epoch": 0.5333333333333333, |
| "grad_norm": 1.0952787399291992, |
| "learning_rate": 8.985645933014356e-06, |
| "loss": 0.2779211699962616, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.32038, |
| "step": 112, |
| "tokens/total": 458752, |
| "tokens/train_per_sec_per_gpu": 8.53, |
| "tokens/trainable": 32977 |
| }, |
| { |
| "epoch": 0.5380952380952381, |
| "grad_norm": 0.8119493126869202, |
| "learning_rate": 8.976076555023924e-06, |
| "loss": 0.16085243225097656, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17451, |
| "step": 113, |
| "tokens/total": 462848, |
| "tokens/train_per_sec_per_gpu": 11.2, |
| "tokens/trainable": 33218 |
| }, |
| { |
| "epoch": 0.5428571428571428, |
| "grad_norm": 0.9444132447242737, |
| "learning_rate": 8.966507177033493e-06, |
| "loss": 0.23129265010356903, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.26023, |
| "step": 114, |
| "tokens/total": 466944, |
| "tokens/train_per_sec_per_gpu": 18.1, |
| "tokens/trainable": 33530 |
| }, |
| { |
| "epoch": 0.5476190476190477, |
| "grad_norm": 0.7226268649101257, |
| "learning_rate": 8.956937799043062e-06, |
| "loss": 0.1263541579246521, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13468, |
| "step": 115, |
| "tokens/total": 471040, |
| "tokens/train_per_sec_per_gpu": 9.2, |
| "tokens/trainable": 33777 |
| }, |
| { |
| "epoch": 0.5523809523809524, |
| "grad_norm": 0.5879023671150208, |
| "learning_rate": 8.947368421052632e-06, |
| "loss": 0.1447187066078186, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15571, |
| "step": 116, |
| "tokens/total": 475136, |
| "tokens/train_per_sec_per_gpu": 18.97, |
| "tokens/trainable": 34086 |
| }, |
| { |
| "epoch": 0.5571428571428572, |
| "grad_norm": 0.6943342089653015, |
| "learning_rate": 8.937799043062201e-06, |
| "loss": 0.21247875690460205, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.23674, |
| "step": 117, |
| "tokens/total": 479232, |
| "tokens/train_per_sec_per_gpu": 21.7, |
| "tokens/trainable": 34575 |
| }, |
| { |
| "epoch": 0.5619047619047619, |
| "grad_norm": 0.5375131368637085, |
| "learning_rate": 8.92822966507177e-06, |
| "loss": 0.11210613697767258, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11863, |
| "step": 118, |
| "tokens/total": 483328, |
| "tokens/train_per_sec_per_gpu": 12.39, |
| "tokens/trainable": 34863 |
| }, |
| { |
| "epoch": 0.5666666666666667, |
| "grad_norm": 0.4493964612483978, |
| "learning_rate": 8.91866028708134e-06, |
| "loss": 0.12835997343063354, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13696, |
| "step": 119, |
| "tokens/total": 487424, |
| "tokens/train_per_sec_per_gpu": 12.58, |
| "tokens/trainable": 35131 |
| }, |
| { |
| "epoch": 0.5714285714285714, |
| "grad_norm": 0.49881812930107117, |
| "learning_rate": 8.90909090909091e-06, |
| "loss": 0.09497980773448944, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09964, |
| "step": 120, |
| "tokens/total": 491520, |
| "tokens/train_per_sec_per_gpu": 11.47, |
| "tokens/trainable": 35390 |
| }, |
| { |
| "epoch": 0.5761904761904761, |
| "grad_norm": 0.3289479911327362, |
| "learning_rate": 8.899521531100479e-06, |
| "loss": 0.10585616528987885, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11166, |
| "step": 121, |
| "tokens/total": 495616, |
| "tokens/train_per_sec_per_gpu": 19.33, |
| "tokens/trainable": 35769 |
| }, |
| { |
| "epoch": 0.580952380952381, |
| "grad_norm": 0.4215591847896576, |
| "learning_rate": 8.889952153110048e-06, |
| "loss": 0.12260347604751587, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13044, |
| "step": 122, |
| "tokens/total": 499712, |
| "tokens/train_per_sec_per_gpu": 13.18, |
| "tokens/trainable": 36093 |
| }, |
| { |
| "epoch": 0.5857142857142857, |
| "grad_norm": 0.6320652961730957, |
| "learning_rate": 8.880382775119618e-06, |
| "loss": 0.19350723922252655, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.2135, |
| "step": 123, |
| "tokens/total": 503808, |
| "tokens/train_per_sec_per_gpu": 14.81, |
| "tokens/trainable": 36465 |
| }, |
| { |
| "epoch": 0.5904761904761905, |
| "grad_norm": 0.7369839549064636, |
| "learning_rate": 8.870813397129187e-06, |
| "loss": 0.2397083193063736, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.27088, |
| "step": 124, |
| "tokens/total": 507904, |
| "tokens/train_per_sec_per_gpu": 18.95, |
| "tokens/trainable": 36788 |
| }, |
| { |
| "epoch": 0.5952380952380952, |
| "grad_norm": 0.5524533987045288, |
| "learning_rate": 8.861244019138756e-06, |
| "loss": 0.1478738933801651, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15937, |
| "step": 125, |
| "tokens/total": 512000, |
| "tokens/train_per_sec_per_gpu": 14.28, |
| "tokens/trainable": 37133 |
| }, |
| { |
| "epoch": 0.6, |
| "grad_norm": 0.7409418821334839, |
| "learning_rate": 8.851674641148326e-06, |
| "loss": 0.2038002759218216, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.22605, |
| "step": 126, |
| "tokens/total": 516096, |
| "tokens/train_per_sec_per_gpu": 13.53, |
| "tokens/trainable": 37417 |
| }, |
| { |
| "epoch": 0.6047619047619047, |
| "grad_norm": 0.8146377801895142, |
| "learning_rate": 8.842105263157895e-06, |
| "loss": 0.22553369402885437, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.25299, |
| "step": 127, |
| "tokens/total": 520192, |
| "tokens/train_per_sec_per_gpu": 9.81, |
| "tokens/trainable": 37702 |
| }, |
| { |
| "epoch": 0.6095238095238096, |
| "grad_norm": 0.6468098163604736, |
| "learning_rate": 8.832535885167465e-06, |
| "loss": 0.16511306166648865, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.17953, |
| "step": 128, |
| "tokens/total": 524288, |
| "tokens/train_per_sec_per_gpu": 20.78, |
| "tokens/trainable": 38034 |
| }, |
| { |
| "epoch": 0.6142857142857143, |
| "grad_norm": 0.8367134928703308, |
| "learning_rate": 8.822966507177034e-06, |
| "loss": 0.11573076993227005, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12269, |
| "step": 129, |
| "tokens/total": 528384, |
| "tokens/train_per_sec_per_gpu": 10.99, |
| "tokens/trainable": 38273 |
| }, |
| { |
| "epoch": 0.6190476190476191, |
| "grad_norm": 0.4852752089500427, |
| "learning_rate": 8.813397129186603e-06, |
| "loss": 0.13677959144115448, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14658, |
| "step": 130, |
| "tokens/total": 532480, |
| "tokens/train_per_sec_per_gpu": 10.06, |
| "tokens/trainable": 38493 |
| }, |
| { |
| "epoch": 0.6238095238095238, |
| "grad_norm": 0.5373315811157227, |
| "learning_rate": 8.803827751196173e-06, |
| "loss": 0.16031202673912048, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17388, |
| "step": 131, |
| "tokens/total": 536576, |
| "tokens/train_per_sec_per_gpu": 13.34, |
| "tokens/trainable": 38840 |
| }, |
| { |
| "epoch": 0.6285714285714286, |
| "grad_norm": 0.9102168679237366, |
| "learning_rate": 8.794258373205742e-06, |
| "loss": 0.30758410692214966, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.36014, |
| "step": 132, |
| "tokens/total": 540672, |
| "tokens/train_per_sec_per_gpu": 15.05, |
| "tokens/trainable": 39169 |
| }, |
| { |
| "epoch": 0.6333333333333333, |
| "grad_norm": 1.0897072553634644, |
| "learning_rate": 8.784688995215311e-06, |
| "loss": 0.21592794358730316, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.24101, |
| "step": 133, |
| "tokens/total": 544768, |
| "tokens/train_per_sec_per_gpu": 8.43, |
| "tokens/trainable": 39424 |
| }, |
| { |
| "epoch": 0.638095238095238, |
| "grad_norm": 0.6847906708717346, |
| "learning_rate": 8.775119617224881e-06, |
| "loss": 0.18241584300994873, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.20011, |
| "step": 134, |
| "tokens/total": 548864, |
| "tokens/train_per_sec_per_gpu": 10.91, |
| "tokens/trainable": 39650 |
| }, |
| { |
| "epoch": 0.6428571428571429, |
| "grad_norm": 0.6738935708999634, |
| "learning_rate": 8.76555023923445e-06, |
| "loss": 0.2037346065044403, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.22597, |
| "step": 135, |
| "tokens/total": 552960, |
| "tokens/train_per_sec_per_gpu": 15.05, |
| "tokens/trainable": 39957 |
| }, |
| { |
| "epoch": 0.6476190476190476, |
| "grad_norm": 1.0508428812026978, |
| "learning_rate": 8.75598086124402e-06, |
| "loss": 0.21129491925239563, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.23528, |
| "step": 136, |
| "tokens/total": 557056, |
| "tokens/train_per_sec_per_gpu": 10.28, |
| "tokens/trainable": 40219 |
| }, |
| { |
| "epoch": 0.6523809523809524, |
| "grad_norm": 0.7576251029968262, |
| "learning_rate": 8.746411483253589e-06, |
| "loss": 0.1137838065624237, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12051, |
| "step": 137, |
| "tokens/total": 561152, |
| "tokens/train_per_sec_per_gpu": 9.86, |
| "tokens/trainable": 40497 |
| }, |
| { |
| "epoch": 0.6571428571428571, |
| "grad_norm": 1.433799147605896, |
| "learning_rate": 8.736842105263158e-06, |
| "loss": 0.31193768978118896, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.36607, |
| "step": 138, |
| "tokens/total": 565248, |
| "tokens/train_per_sec_per_gpu": 15.47, |
| "tokens/trainable": 40802 |
| }, |
| { |
| "epoch": 0.6619047619047619, |
| "grad_norm": 0.9841307997703552, |
| "learning_rate": 8.727272727272728e-06, |
| "loss": 0.17987215518951416, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.19706, |
| "step": 139, |
| "tokens/total": 569344, |
| "tokens/train_per_sec_per_gpu": 9.93, |
| "tokens/trainable": 41048 |
| }, |
| { |
| "epoch": 0.6666666666666666, |
| "grad_norm": 0.7832286357879639, |
| "learning_rate": 8.717703349282297e-06, |
| "loss": 0.21015842258930206, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.23387, |
| "step": 140, |
| "tokens/total": 573440, |
| "tokens/train_per_sec_per_gpu": 18.11, |
| "tokens/trainable": 41428 |
| }, |
| { |
| "epoch": 0.6714285714285714, |
| "grad_norm": 0.8383325934410095, |
| "learning_rate": 8.708133971291867e-06, |
| "loss": 0.21185638010501862, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.23597, |
| "step": 141, |
| "tokens/total": 577536, |
| "tokens/train_per_sec_per_gpu": 14.71, |
| "tokens/trainable": 41744 |
| }, |
| { |
| "epoch": 0.6761904761904762, |
| "grad_norm": 0.680001437664032, |
| "learning_rate": 8.698564593301436e-06, |
| "loss": 0.21891218423843384, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.24472, |
| "step": 142, |
| "tokens/total": 581632, |
| "tokens/train_per_sec_per_gpu": 16.03, |
| "tokens/trainable": 42065 |
| }, |
| { |
| "epoch": 0.680952380952381, |
| "grad_norm": 0.6654219031333923, |
| "learning_rate": 8.688995215311005e-06, |
| "loss": 0.16606123745441437, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.18065, |
| "step": 143, |
| "tokens/total": 585728, |
| "tokens/train_per_sec_per_gpu": 16.31, |
| "tokens/trainable": 42389 |
| }, |
| { |
| "epoch": 0.6857142857142857, |
| "grad_norm": 0.9814354181289673, |
| "learning_rate": 8.679425837320575e-06, |
| "loss": 0.19918571412563324, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.22041, |
| "step": 144, |
| "tokens/total": 589824, |
| "tokens/train_per_sec_per_gpu": 22.97, |
| "tokens/trainable": 42791 |
| }, |
| { |
| "epoch": 0.6904761904761905, |
| "grad_norm": 0.7615348100662231, |
| "learning_rate": 8.669856459330144e-06, |
| "loss": 0.244194895029068, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.27659, |
| "step": 145, |
| "tokens/total": 593920, |
| "tokens/train_per_sec_per_gpu": 9.59, |
| "tokens/trainable": 43054 |
| }, |
| { |
| "epoch": 0.6952380952380952, |
| "grad_norm": 0.6813610196113586, |
| "learning_rate": 8.660287081339714e-06, |
| "loss": 0.21728792786598206, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.2427, |
| "step": 146, |
| "tokens/total": 598016, |
| "tokens/train_per_sec_per_gpu": 11.93, |
| "tokens/trainable": 43332 |
| }, |
| { |
| "epoch": 0.7, |
| "grad_norm": 0.7670934796333313, |
| "learning_rate": 8.650717703349283e-06, |
| "loss": 0.224639892578125, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.25187, |
| "step": 147, |
| "tokens/total": 602112, |
| "tokens/train_per_sec_per_gpu": 15.52, |
| "tokens/trainable": 43610 |
| }, |
| { |
| "epoch": 0.7047619047619048, |
| "grad_norm": 0.4526863992214203, |
| "learning_rate": 8.641148325358852e-06, |
| "loss": 0.14494425058364868, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.15598, |
| "step": 148, |
| "tokens/total": 606208, |
| "tokens/train_per_sec_per_gpu": 27.26, |
| "tokens/trainable": 44032 |
| }, |
| { |
| "epoch": 0.7095238095238096, |
| "grad_norm": 0.8980340957641602, |
| "learning_rate": 8.631578947368422e-06, |
| "loss": 0.23772993683815002, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.26837, |
| "step": 149, |
| "tokens/total": 610304, |
| "tokens/train_per_sec_per_gpu": 11.17, |
| "tokens/trainable": 44233 |
| }, |
| { |
| "epoch": 0.7142857142857143, |
| "grad_norm": 0.8445303440093994, |
| "learning_rate": 8.622009569377991e-06, |
| "loss": 0.21160978078842163, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.23567, |
| "step": 150, |
| "tokens/total": 614400, |
| "tokens/train_per_sec_per_gpu": 13.91, |
| "tokens/trainable": 44535 |
| }, |
| { |
| "epoch": 0.719047619047619, |
| "grad_norm": 0.9266573786735535, |
| "learning_rate": 8.61244019138756e-06, |
| "loss": 0.09477423876523972, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09941, |
| "step": 151, |
| "tokens/total": 618496, |
| "tokens/train_per_sec_per_gpu": 9.47, |
| "tokens/trainable": 44780 |
| }, |
| { |
| "epoch": 0.7238095238095238, |
| "grad_norm": 0.7422537803649902, |
| "learning_rate": 8.60287081339713e-06, |
| "loss": 0.12581270933151245, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13407, |
| "step": 152, |
| "tokens/total": 622592, |
| "tokens/train_per_sec_per_gpu": 10.39, |
| "tokens/trainable": 45041 |
| }, |
| { |
| "epoch": 0.7285714285714285, |
| "grad_norm": 0.709301233291626, |
| "learning_rate": 8.5933014354067e-06, |
| "loss": 0.1559009552001953, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16871, |
| "step": 153, |
| "tokens/total": 626688, |
| "tokens/train_per_sec_per_gpu": 16.34, |
| "tokens/trainable": 45354 |
| }, |
| { |
| "epoch": 0.7333333333333333, |
| "grad_norm": 0.710270345211029, |
| "learning_rate": 8.583732057416269e-06, |
| "loss": 0.16102313995361328, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17471, |
| "step": 154, |
| "tokens/total": 630784, |
| "tokens/train_per_sec_per_gpu": 13.98, |
| "tokens/trainable": 45679 |
| }, |
| { |
| "epoch": 0.7380952380952381, |
| "grad_norm": 0.5672938823699951, |
| "learning_rate": 8.574162679425838e-06, |
| "loss": 0.14538349211215973, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15648, |
| "step": 155, |
| "tokens/total": 634880, |
| "tokens/train_per_sec_per_gpu": 13.37, |
| "tokens/trainable": 46000 |
| }, |
| { |
| "epoch": 0.7428571428571429, |
| "grad_norm": 0.5063062906265259, |
| "learning_rate": 8.564593301435407e-06, |
| "loss": 0.1216127872467041, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12932, |
| "step": 156, |
| "tokens/total": 638976, |
| "tokens/train_per_sec_per_gpu": 15.37, |
| "tokens/trainable": 46248 |
| }, |
| { |
| "epoch": 0.7476190476190476, |
| "grad_norm": 0.8823851346969604, |
| "learning_rate": 8.555023923444977e-06, |
| "loss": 0.20955757796764374, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.23313, |
| "step": 157, |
| "tokens/total": 643072, |
| "tokens/train_per_sec_per_gpu": 15.44, |
| "tokens/trainable": 46557 |
| }, |
| { |
| "epoch": 0.7523809523809524, |
| "grad_norm": 0.6821964979171753, |
| "learning_rate": 8.545454545454546e-06, |
| "loss": 0.12059423327445984, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12817, |
| "step": 158, |
| "tokens/total": 647168, |
| "tokens/train_per_sec_per_gpu": 14.4, |
| "tokens/trainable": 46839 |
| }, |
| { |
| "epoch": 0.7571428571428571, |
| "grad_norm": 1.7155925035476685, |
| "learning_rate": 8.535885167464116e-06, |
| "loss": 0.16679029166698456, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.18151, |
| "step": 159, |
| "tokens/total": 651264, |
| "tokens/train_per_sec_per_gpu": 7.54, |
| "tokens/trainable": 47047 |
| }, |
| { |
| "epoch": 0.7619047619047619, |
| "grad_norm": 0.7815935015678406, |
| "learning_rate": 8.526315789473685e-06, |
| "loss": 0.20110413432121277, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.22275, |
| "step": 160, |
| "tokens/total": 655360, |
| "tokens/train_per_sec_per_gpu": 11.91, |
| "tokens/trainable": 47349 |
| }, |
| { |
| "epoch": 0.7666666666666667, |
| "grad_norm": 0.8044136762619019, |
| "learning_rate": 8.516746411483254e-06, |
| "loss": 0.1733863651752472, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.18933, |
| "step": 161, |
| "tokens/total": 659456, |
| "tokens/train_per_sec_per_gpu": 13.39, |
| "tokens/trainable": 47685 |
| }, |
| { |
| "epoch": 0.7714285714285715, |
| "grad_norm": 0.6603718400001526, |
| "learning_rate": 8.507177033492824e-06, |
| "loss": 0.13823741674423218, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14825, |
| "step": 162, |
| "tokens/total": 663552, |
| "tokens/train_per_sec_per_gpu": 13.32, |
| "tokens/trainable": 47978 |
| }, |
| { |
| "epoch": 0.7761904761904762, |
| "grad_norm": 0.7813670039176941, |
| "learning_rate": 8.497607655502393e-06, |
| "loss": 0.19184273481369019, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.21148, |
| "step": 163, |
| "tokens/total": 667648, |
| "tokens/train_per_sec_per_gpu": 14.04, |
| "tokens/trainable": 48239 |
| }, |
| { |
| "epoch": 0.780952380952381, |
| "grad_norm": 0.6289886832237244, |
| "learning_rate": 8.488038277511963e-06, |
| "loss": 0.12562716007232666, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13386, |
| "step": 164, |
| "tokens/total": 671744, |
| "tokens/train_per_sec_per_gpu": 9.43, |
| "tokens/trainable": 48462 |
| }, |
| { |
| "epoch": 0.7857142857142857, |
| "grad_norm": 0.7649319767951965, |
| "learning_rate": 8.478468899521532e-06, |
| "loss": 0.2198971062898636, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.24595, |
| "step": 165, |
| "tokens/total": 675840, |
| "tokens/train_per_sec_per_gpu": 14.98, |
| "tokens/trainable": 48758 |
| }, |
| { |
| "epoch": 0.7904761904761904, |
| "grad_norm": 0.793499231338501, |
| "learning_rate": 8.468899521531101e-06, |
| "loss": 0.1740531027317047, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.19012, |
| "step": 166, |
| "tokens/total": 679936, |
| "tokens/train_per_sec_per_gpu": 12.74, |
| "tokens/trainable": 49063 |
| }, |
| { |
| "epoch": 0.7952380952380952, |
| "grad_norm": 0.6194958090782166, |
| "learning_rate": 8.45933014354067e-06, |
| "loss": 0.11570701003074646, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12267, |
| "step": 167, |
| "tokens/total": 684032, |
| "tokens/train_per_sec_per_gpu": 12.72, |
| "tokens/trainable": 49304 |
| }, |
| { |
| "epoch": 0.8, |
| "grad_norm": 0.46994757652282715, |
| "learning_rate": 8.44976076555024e-06, |
| "loss": 0.09198522567749023, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09635, |
| "step": 168, |
| "tokens/total": 688128, |
| "tokens/train_per_sec_per_gpu": 16.65, |
| "tokens/trainable": 49632 |
| }, |
| { |
| "epoch": 0.8047619047619048, |
| "grad_norm": 0.7131854295730591, |
| "learning_rate": 8.44019138755981e-06, |
| "loss": 0.16379094123840332, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17797, |
| "step": 169, |
| "tokens/total": 692224, |
| "tokens/train_per_sec_per_gpu": 11.06, |
| "tokens/trainable": 49940 |
| }, |
| { |
| "epoch": 0.8095238095238095, |
| "grad_norm": 0.6772812008857727, |
| "learning_rate": 8.430622009569379e-06, |
| "loss": 0.13202588260173798, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14114, |
| "step": 170, |
| "tokens/total": 696320, |
| "tokens/train_per_sec_per_gpu": 9.54, |
| "tokens/trainable": 50179 |
| }, |
| { |
| "epoch": 0.8142857142857143, |
| "grad_norm": 0.8214194774627686, |
| "learning_rate": 8.421052631578948e-06, |
| "loss": 0.19660362601280212, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.21726, |
| "step": 171, |
| "tokens/total": 700416, |
| "tokens/train_per_sec_per_gpu": 15.79, |
| "tokens/trainable": 50476 |
| }, |
| { |
| "epoch": 0.819047619047619, |
| "grad_norm": 0.6981196403503418, |
| "learning_rate": 8.411483253588518e-06, |
| "loss": 0.1562727838754654, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.16915, |
| "step": 172, |
| "tokens/total": 704512, |
| "tokens/train_per_sec_per_gpu": 10.79, |
| "tokens/trainable": 50755 |
| }, |
| { |
| "epoch": 0.8238095238095238, |
| "grad_norm": 0.884662389755249, |
| "learning_rate": 8.401913875598087e-06, |
| "loss": 0.19947463274002075, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.22076, |
| "step": 173, |
| "tokens/total": 708608, |
| "tokens/train_per_sec_per_gpu": 11.33, |
| "tokens/trainable": 51003 |
| }, |
| { |
| "epoch": 0.8285714285714286, |
| "grad_norm": 0.7139277458190918, |
| "learning_rate": 8.392344497607656e-06, |
| "loss": 0.19828957319259644, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.21932, |
| "step": 174, |
| "tokens/total": 712704, |
| "tokens/train_per_sec_per_gpu": 11.04, |
| "tokens/trainable": 51257 |
| }, |
| { |
| "epoch": 0.8333333333333334, |
| "grad_norm": 0.5720876455307007, |
| "learning_rate": 8.382775119617226e-06, |
| "loss": 0.11307372152805328, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11971, |
| "step": 175, |
| "tokens/total": 716800, |
| "tokens/train_per_sec_per_gpu": 10.63, |
| "tokens/trainable": 51499 |
| }, |
| { |
| "epoch": 0.8380952380952381, |
| "grad_norm": 0.8213474154472351, |
| "learning_rate": 8.373205741626795e-06, |
| "loss": 0.21523649990558624, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.24016, |
| "step": 176, |
| "tokens/total": 720896, |
| "tokens/train_per_sec_per_gpu": 17.26, |
| "tokens/trainable": 51805 |
| }, |
| { |
| "epoch": 0.8428571428571429, |
| "grad_norm": 0.5487144589424133, |
| "learning_rate": 8.363636363636365e-06, |
| "loss": 0.14880937337875366, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16045, |
| "step": 177, |
| "tokens/total": 724992, |
| "tokens/train_per_sec_per_gpu": 11.26, |
| "tokens/trainable": 52127 |
| }, |
| { |
| "epoch": 0.8476190476190476, |
| "grad_norm": 1.5928086042404175, |
| "learning_rate": 8.354066985645934e-06, |
| "loss": 0.18438681960105896, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.20248, |
| "step": 178, |
| "tokens/total": 729088, |
| "tokens/train_per_sec_per_gpu": 10.2, |
| "tokens/trainable": 52374 |
| }, |
| { |
| "epoch": 0.8523809523809524, |
| "grad_norm": 0.872934103012085, |
| "learning_rate": 8.344497607655503e-06, |
| "loss": 0.15895763039588928, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.17229, |
| "step": 179, |
| "tokens/total": 733184, |
| "tokens/train_per_sec_per_gpu": 10.91, |
| "tokens/trainable": 52612 |
| }, |
| { |
| "epoch": 0.8571428571428571, |
| "grad_norm": 0.7492268085479736, |
| "learning_rate": 8.334928229665073e-06, |
| "loss": 0.19397899508476257, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.21407, |
| "step": 180, |
| "tokens/total": 737280, |
| "tokens/train_per_sec_per_gpu": 11.08, |
| "tokens/trainable": 52935 |
| }, |
| { |
| "epoch": 0.861904761904762, |
| "grad_norm": 1.050233244895935, |
| "learning_rate": 8.325358851674642e-06, |
| "loss": 0.09477034956216812, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09941, |
| "step": 181, |
| "tokens/total": 741376, |
| "tokens/train_per_sec_per_gpu": 9.84, |
| "tokens/trainable": 53220 |
| }, |
| { |
| "epoch": 0.8666666666666667, |
| "grad_norm": 0.4782058894634247, |
| "learning_rate": 8.315789473684212e-06, |
| "loss": 0.11910860240459442, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12649, |
| "step": 182, |
| "tokens/total": 745472, |
| "tokens/train_per_sec_per_gpu": 16.75, |
| "tokens/trainable": 53535 |
| }, |
| { |
| "epoch": 0.8714285714285714, |
| "grad_norm": 0.906046986579895, |
| "learning_rate": 8.306220095693781e-06, |
| "loss": 0.2756553590297699, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.31739, |
| "step": 183, |
| "tokens/total": 749568, |
| "tokens/train_per_sec_per_gpu": 9.67, |
| "tokens/trainable": 53797 |
| }, |
| { |
| "epoch": 0.8761904761904762, |
| "grad_norm": 0.8429417014122009, |
| "learning_rate": 8.29665071770335e-06, |
| "loss": 0.15779748558998108, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17093, |
| "step": 184, |
| "tokens/total": 753664, |
| "tokens/train_per_sec_per_gpu": 24.29, |
| "tokens/trainable": 54194 |
| }, |
| { |
| "epoch": 0.8809523809523809, |
| "grad_norm": 1.0448087453842163, |
| "learning_rate": 8.28708133971292e-06, |
| "loss": 0.1565137505531311, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16943, |
| "step": 185, |
| "tokens/total": 757760, |
| "tokens/train_per_sec_per_gpu": 10.63, |
| "tokens/trainable": 54403 |
| }, |
| { |
| "epoch": 0.8857142857142857, |
| "grad_norm": 0.760995626449585, |
| "learning_rate": 8.277511961722489e-06, |
| "loss": 0.21484459936618805, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.23967, |
| "step": 186, |
| "tokens/total": 761856, |
| "tokens/train_per_sec_per_gpu": 22.28, |
| "tokens/trainable": 54744 |
| }, |
| { |
| "epoch": 0.8904761904761904, |
| "grad_norm": 0.6110489368438721, |
| "learning_rate": 8.267942583732058e-06, |
| "loss": 0.15965917706489563, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17311, |
| "step": 187, |
| "tokens/total": 765952, |
| "tokens/train_per_sec_per_gpu": 13.98, |
| "tokens/trainable": 55043 |
| }, |
| { |
| "epoch": 0.8952380952380953, |
| "grad_norm": 0.6338987350463867, |
| "learning_rate": 8.258373205741628e-06, |
| "loss": 0.1492011845111847, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.16091, |
| "step": 188, |
| "tokens/total": 770048, |
| "tokens/train_per_sec_per_gpu": 13.88, |
| "tokens/trainable": 55329 |
| }, |
| { |
| "epoch": 0.9, |
| "grad_norm": 0.6387603878974915, |
| "learning_rate": 8.248803827751197e-06, |
| "loss": 0.1357661634683609, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14541, |
| "step": 189, |
| "tokens/total": 774144, |
| "tokens/train_per_sec_per_gpu": 11.66, |
| "tokens/trainable": 55601 |
| }, |
| { |
| "epoch": 0.9047619047619048, |
| "grad_norm": 0.49371299147605896, |
| "learning_rate": 8.239234449760765e-06, |
| "loss": 0.12431974709033966, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13238, |
| "step": 190, |
| "tokens/total": 778240, |
| "tokens/train_per_sec_per_gpu": 10.34, |
| "tokens/trainable": 55919 |
| }, |
| { |
| "epoch": 0.9095238095238095, |
| "grad_norm": 0.6698276400566101, |
| "learning_rate": 8.229665071770336e-06, |
| "loss": 0.1634908765554428, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.17761, |
| "step": 191, |
| "tokens/total": 782336, |
| "tokens/train_per_sec_per_gpu": 13.4, |
| "tokens/trainable": 56188 |
| }, |
| { |
| "epoch": 0.9142857142857143, |
| "grad_norm": 0.8588296175003052, |
| "learning_rate": 8.220095693779905e-06, |
| "loss": 0.12147596478462219, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12916, |
| "step": 192, |
| "tokens/total": 786432, |
| "tokens/train_per_sec_per_gpu": 10.73, |
| "tokens/trainable": 56440 |
| }, |
| { |
| "epoch": 0.919047619047619, |
| "grad_norm": 0.742889940738678, |
| "learning_rate": 8.210526315789475e-06, |
| "loss": 0.11071045696735382, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11707, |
| "step": 193, |
| "tokens/total": 790528, |
| "tokens/train_per_sec_per_gpu": 17.25, |
| "tokens/trainable": 56765 |
| }, |
| { |
| "epoch": 0.9238095238095239, |
| "grad_norm": 0.6750599145889282, |
| "learning_rate": 8.200956937799044e-06, |
| "loss": 0.1338326781988144, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1432, |
| "step": 194, |
| "tokens/total": 794624, |
| "tokens/train_per_sec_per_gpu": 13.36, |
| "tokens/trainable": 57089 |
| }, |
| { |
| "epoch": 0.9285714285714286, |
| "grad_norm": 0.6421961784362793, |
| "learning_rate": 8.191387559808614e-06, |
| "loss": 0.14603616297245026, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15724, |
| "step": 195, |
| "tokens/total": 798720, |
| "tokens/train_per_sec_per_gpu": 12.22, |
| "tokens/trainable": 57404 |
| }, |
| { |
| "epoch": 0.9333333333333333, |
| "grad_norm": 1.3169405460357666, |
| "learning_rate": 8.181818181818183e-06, |
| "loss": 0.1660291999578476, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.18061, |
| "step": 196, |
| "tokens/total": 802816, |
| "tokens/train_per_sec_per_gpu": 8.48, |
| "tokens/trainable": 57637 |
| }, |
| { |
| "epoch": 0.9380952380952381, |
| "grad_norm": 0.6492071151733398, |
| "learning_rate": 8.17224880382775e-06, |
| "loss": 0.16259926557540894, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17657, |
| "step": 197, |
| "tokens/total": 806912, |
| "tokens/train_per_sec_per_gpu": 14.91, |
| "tokens/trainable": 58007 |
| }, |
| { |
| "epoch": 0.9428571428571428, |
| "grad_norm": 0.6238608956336975, |
| "learning_rate": 8.162679425837322e-06, |
| "loss": 0.1267922967672348, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13518, |
| "step": 198, |
| "tokens/total": 811008, |
| "tokens/train_per_sec_per_gpu": 17.72, |
| "tokens/trainable": 58294 |
| }, |
| { |
| "epoch": 0.9476190476190476, |
| "grad_norm": 0.8982058167457581, |
| "learning_rate": 8.153110047846891e-06, |
| "loss": 0.11205602437257767, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11858, |
| "step": 199, |
| "tokens/total": 815104, |
| "tokens/train_per_sec_per_gpu": 14.64, |
| "tokens/trainable": 58635 |
| }, |
| { |
| "epoch": 0.9523809523809523, |
| "grad_norm": 0.6866769790649414, |
| "learning_rate": 8.14354066985646e-06, |
| "loss": 0.18688730895519257, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.20549, |
| "step": 200, |
| "tokens/total": 819200, |
| "tokens/train_per_sec_per_gpu": 15.36, |
| "tokens/trainable": 58983 |
| }, |
| { |
| "epoch": 0.9571428571428572, |
| "grad_norm": 0.7195040583610535, |
| "learning_rate": 8.13397129186603e-06, |
| "loss": 0.14748433232307434, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15892, |
| "step": 201, |
| "tokens/total": 823296, |
| "tokens/train_per_sec_per_gpu": 9.98, |
| "tokens/trainable": 59225 |
| }, |
| { |
| "epoch": 0.9619047619047619, |
| "grad_norm": 0.5893474817276001, |
| "learning_rate": 8.1244019138756e-06, |
| "loss": 0.11337488144636154, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12005, |
| "step": 202, |
| "tokens/total": 827392, |
| "tokens/train_per_sec_per_gpu": 14.95, |
| "tokens/trainable": 59527 |
| }, |
| { |
| "epoch": 0.9666666666666667, |
| "grad_norm": 0.8447104096412659, |
| "learning_rate": 8.114832535885167e-06, |
| "loss": 0.14601489901542664, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15721, |
| "step": 203, |
| "tokens/total": 831488, |
| "tokens/train_per_sec_per_gpu": 13.82, |
| "tokens/trainable": 59774 |
| }, |
| { |
| "epoch": 0.9714285714285714, |
| "grad_norm": 0.7004845142364502, |
| "learning_rate": 8.105263157894736e-06, |
| "loss": 0.1523507833480835, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16457, |
| "step": 204, |
| "tokens/total": 835584, |
| "tokens/train_per_sec_per_gpu": 9.08, |
| "tokens/trainable": 60017 |
| }, |
| { |
| "epoch": 0.9761904761904762, |
| "grad_norm": 0.8089159727096558, |
| "learning_rate": 8.095693779904307e-06, |
| "loss": 0.07807943224906921, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08121, |
| "step": 205, |
| "tokens/total": 839680, |
| "tokens/train_per_sec_per_gpu": 9.57, |
| "tokens/trainable": 60304 |
| }, |
| { |
| "epoch": 0.9809523809523809, |
| "grad_norm": 1.0327520370483398, |
| "learning_rate": 8.086124401913877e-06, |
| "loss": 0.14542421698570251, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15653, |
| "step": 206, |
| "tokens/total": 843776, |
| "tokens/train_per_sec_per_gpu": 12.38, |
| "tokens/trainable": 60538 |
| }, |
| { |
| "epoch": 0.9857142857142858, |
| "grad_norm": 0.551845371723175, |
| "learning_rate": 8.076555023923446e-06, |
| "loss": 0.10232502222061157, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.10774, |
| "step": 207, |
| "tokens/total": 847872, |
| "tokens/train_per_sec_per_gpu": 11.11, |
| "tokens/trainable": 60799 |
| }, |
| { |
| "epoch": 0.9904761904761905, |
| "grad_norm": 0.7714454531669617, |
| "learning_rate": 8.066985645933016e-06, |
| "loss": 0.14636501669883728, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15762, |
| "step": 208, |
| "tokens/total": 851968, |
| "tokens/train_per_sec_per_gpu": 11.33, |
| "tokens/trainable": 61041 |
| }, |
| { |
| "epoch": 0.9952380952380953, |
| "grad_norm": 0.6568189859390259, |
| "learning_rate": 8.057416267942585e-06, |
| "loss": 0.12750931084156036, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.136, |
| "step": 209, |
| "tokens/total": 856064, |
| "tokens/train_per_sec_per_gpu": 12.49, |
| "tokens/trainable": 61298 |
| }, |
| { |
| "epoch": 1.0, |
| "grad_norm": 0.532839298248291, |
| "learning_rate": 8.047846889952153e-06, |
| "loss": 0.09202240407466888, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09639, |
| "step": 210, |
| "tokens/total": 860160, |
| "tokens/train_per_sec_per_gpu": 8.6, |
| "tokens/trainable": 61610 |
| }, |
| { |
| "epoch": 1.0, |
| "eval_loss": 0.13460691273212433, |
| "eval_ppl": 1.14409, |
| "eval_runtime": 10.3792, |
| "eval_samples_per_second": 8.575, |
| "eval_steps_per_second": 2.216, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 35.58, |
| "memory/max_allocated (GiB)": 35.58, |
| "step": 210 |
| }, |
| { |
| "epoch": 1.0047619047619047, |
| "grad_norm": 0.6392420530319214, |
| "learning_rate": 8.038277511961722e-06, |
| "loss": 0.10671957582235336, |
| "memory/device_reserved (GiB)": 39.15, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11262, |
| "step": 211, |
| "tokens/total": 864256, |
| "tokens/train_per_sec_per_gpu": 11.9, |
| "tokens/trainable": 61889 |
| }, |
| { |
| "epoch": 1.0095238095238095, |
| "grad_norm": 0.5944091081619263, |
| "learning_rate": 8.028708133971293e-06, |
| "loss": 0.12425249814987183, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1323, |
| "step": 212, |
| "tokens/total": 868352, |
| "tokens/train_per_sec_per_gpu": 17.77, |
| "tokens/trainable": 62213 |
| }, |
| { |
| "epoch": 1.0142857142857142, |
| "grad_norm": 0.7566636800765991, |
| "learning_rate": 8.019138755980863e-06, |
| "loss": 0.1521780788898468, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.16437, |
| "step": 213, |
| "tokens/total": 872448, |
| "tokens/train_per_sec_per_gpu": 14.35, |
| "tokens/trainable": 62580 |
| }, |
| { |
| "epoch": 1.019047619047619, |
| "grad_norm": 0.7836124300956726, |
| "learning_rate": 8.009569377990432e-06, |
| "loss": 0.16885873675346375, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.18395, |
| "step": 214, |
| "tokens/total": 876544, |
| "tokens/train_per_sec_per_gpu": 8.6, |
| "tokens/trainable": 62834 |
| }, |
| { |
| "epoch": 1.0238095238095237, |
| "grad_norm": 0.6088077425956726, |
| "learning_rate": 8.000000000000001e-06, |
| "loss": 0.10565628111362457, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11144, |
| "step": 215, |
| "tokens/total": 880640, |
| "tokens/train_per_sec_per_gpu": 11.49, |
| "tokens/trainable": 63092 |
| }, |
| { |
| "epoch": 1.0285714285714285, |
| "grad_norm": 0.6194921731948853, |
| "learning_rate": 7.990430622009569e-06, |
| "loss": 0.09690128266811371, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10175, |
| "step": 216, |
| "tokens/total": 884736, |
| "tokens/train_per_sec_per_gpu": 9.38, |
| "tokens/trainable": 63383 |
| }, |
| { |
| "epoch": 1.0333333333333334, |
| "grad_norm": 0.6402537226676941, |
| "learning_rate": 7.980861244019138e-06, |
| "loss": 0.11445532739162445, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12126, |
| "step": 217, |
| "tokens/total": 888832, |
| "tokens/train_per_sec_per_gpu": 21.21, |
| "tokens/trainable": 63780 |
| }, |
| { |
| "epoch": 1.0380952380952382, |
| "grad_norm": 0.5895891785621643, |
| "learning_rate": 7.971291866028708e-06, |
| "loss": 0.12593643367290497, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13421, |
| "step": 218, |
| "tokens/total": 892928, |
| "tokens/train_per_sec_per_gpu": 13.17, |
| "tokens/trainable": 64069 |
| }, |
| { |
| "epoch": 1.042857142857143, |
| "grad_norm": 0.7168067097663879, |
| "learning_rate": 7.961722488038279e-06, |
| "loss": 0.14395299553871155, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15483, |
| "step": 219, |
| "tokens/total": 897024, |
| "tokens/train_per_sec_per_gpu": 13.16, |
| "tokens/trainable": 64317 |
| }, |
| { |
| "epoch": 1.0476190476190477, |
| "grad_norm": 0.9142693877220154, |
| "learning_rate": 7.952153110047848e-06, |
| "loss": 0.10631201416254044, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11217, |
| "step": 220, |
| "tokens/total": 901120, |
| "tokens/train_per_sec_per_gpu": 19.77, |
| "tokens/trainable": 64698 |
| }, |
| { |
| "epoch": 1.0523809523809524, |
| "grad_norm": 0.5825505256652832, |
| "learning_rate": 7.942583732057418e-06, |
| "loss": 0.10505223274230957, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11077, |
| "step": 221, |
| "tokens/total": 905216, |
| "tokens/train_per_sec_per_gpu": 11.69, |
| "tokens/trainable": 64991 |
| }, |
| { |
| "epoch": 1.0571428571428572, |
| "grad_norm": 0.5183160901069641, |
| "learning_rate": 7.933014354066987e-06, |
| "loss": 0.08044271171092987, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08377, |
| "step": 222, |
| "tokens/total": 909312, |
| "tokens/train_per_sec_per_gpu": 11.2, |
| "tokens/trainable": 65202 |
| }, |
| { |
| "epoch": 1.061904761904762, |
| "grad_norm": 0.5459516048431396, |
| "learning_rate": 7.923444976076555e-06, |
| "loss": 0.0791759341955185, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.08239, |
| "step": 223, |
| "tokens/total": 913408, |
| "tokens/train_per_sec_per_gpu": 15.83, |
| "tokens/trainable": 65524 |
| }, |
| { |
| "epoch": 1.0666666666666667, |
| "grad_norm": 0.767754852771759, |
| "learning_rate": 7.913875598086124e-06, |
| "loss": 0.13113239407539368, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14012, |
| "step": 224, |
| "tokens/total": 917504, |
| "tokens/train_per_sec_per_gpu": 21.59, |
| "tokens/trainable": 65856 |
| }, |
| { |
| "epoch": 1.0714285714285714, |
| "grad_norm": 0.645051121711731, |
| "learning_rate": 7.904306220095695e-06, |
| "loss": 0.14703069627285004, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15839, |
| "step": 225, |
| "tokens/total": 921600, |
| "tokens/train_per_sec_per_gpu": 13.94, |
| "tokens/trainable": 66181 |
| }, |
| { |
| "epoch": 1.0761904761904761, |
| "grad_norm": 0.4971192181110382, |
| "learning_rate": 7.894736842105265e-06, |
| "loss": 0.08223872631788254, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08571, |
| "step": 226, |
| "tokens/total": 925696, |
| "tokens/train_per_sec_per_gpu": 11.49, |
| "tokens/trainable": 66466 |
| }, |
| { |
| "epoch": 1.0809523809523809, |
| "grad_norm": 0.34934762120246887, |
| "learning_rate": 7.885167464114834e-06, |
| "loss": 0.08169053494930267, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08512, |
| "step": 227, |
| "tokens/total": 929792, |
| "tokens/train_per_sec_per_gpu": 12.48, |
| "tokens/trainable": 66824 |
| }, |
| { |
| "epoch": 1.0857142857142856, |
| "grad_norm": 0.5710490345954895, |
| "learning_rate": 7.875598086124403e-06, |
| "loss": 0.11047933995723724, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11681, |
| "step": 228, |
| "tokens/total": 933888, |
| "tokens/train_per_sec_per_gpu": 20.36, |
| "tokens/trainable": 67193 |
| }, |
| { |
| "epoch": 1.0904761904761904, |
| "grad_norm": 0.6241896748542786, |
| "learning_rate": 7.866028708133971e-06, |
| "loss": 0.08586344122886658, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08966, |
| "step": 229, |
| "tokens/total": 937984, |
| "tokens/train_per_sec_per_gpu": 15.51, |
| "tokens/trainable": 67473 |
| }, |
| { |
| "epoch": 1.0952380952380953, |
| "grad_norm": 0.594308614730835, |
| "learning_rate": 7.85645933014354e-06, |
| "loss": 0.11477260291576385, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12162, |
| "step": 230, |
| "tokens/total": 942080, |
| "tokens/train_per_sec_per_gpu": 13.98, |
| "tokens/trainable": 67737 |
| }, |
| { |
| "epoch": 1.1, |
| "grad_norm": 0.5714224576950073, |
| "learning_rate": 7.84688995215311e-06, |
| "loss": 0.12829524278640747, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13689, |
| "step": 231, |
| "tokens/total": 946176, |
| "tokens/train_per_sec_per_gpu": 20.84, |
| "tokens/trainable": 68052 |
| }, |
| { |
| "epoch": 1.1047619047619048, |
| "grad_norm": 0.8406590819358826, |
| "learning_rate": 7.837320574162681e-06, |
| "loss": 0.1560124307870865, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16884, |
| "step": 232, |
| "tokens/total": 950272, |
| "tokens/train_per_sec_per_gpu": 10.72, |
| "tokens/trainable": 68285 |
| }, |
| { |
| "epoch": 1.1095238095238096, |
| "grad_norm": 0.7849947810173035, |
| "learning_rate": 7.82775119617225e-06, |
| "loss": 0.14618279039859772, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15741, |
| "step": 233, |
| "tokens/total": 954368, |
| "tokens/train_per_sec_per_gpu": 10.91, |
| "tokens/trainable": 68562 |
| }, |
| { |
| "epoch": 1.1142857142857143, |
| "grad_norm": 0.6889620423316956, |
| "learning_rate": 7.81818181818182e-06, |
| "loss": 0.13402847945690155, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14343, |
| "step": 234, |
| "tokens/total": 958464, |
| "tokens/train_per_sec_per_gpu": 13.99, |
| "tokens/trainable": 68995 |
| }, |
| { |
| "epoch": 1.119047619047619, |
| "grad_norm": 0.5359621644020081, |
| "learning_rate": 7.808612440191387e-06, |
| "loss": 0.09991563856601715, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10508, |
| "step": 235, |
| "tokens/total": 962560, |
| "tokens/train_per_sec_per_gpu": 12.45, |
| "tokens/trainable": 69277 |
| }, |
| { |
| "epoch": 1.1238095238095238, |
| "grad_norm": 0.8721844553947449, |
| "learning_rate": 7.799043062200957e-06, |
| "loss": 0.11624078452587128, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12327, |
| "step": 236, |
| "tokens/total": 966656, |
| "tokens/train_per_sec_per_gpu": 15.46, |
| "tokens/trainable": 69547 |
| }, |
| { |
| "epoch": 1.1285714285714286, |
| "grad_norm": 0.5049797892570496, |
| "learning_rate": 7.789473684210526e-06, |
| "loss": 0.10903551429510117, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1152, |
| "step": 237, |
| "tokens/total": 970752, |
| "tokens/train_per_sec_per_gpu": 13.01, |
| "tokens/trainable": 69831 |
| }, |
| { |
| "epoch": 1.1333333333333333, |
| "grad_norm": 0.6197219491004944, |
| "learning_rate": 7.779904306220096e-06, |
| "loss": 0.1088065505027771, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11495, |
| "step": 238, |
| "tokens/total": 974848, |
| "tokens/train_per_sec_per_gpu": 10.43, |
| "tokens/trainable": 70063 |
| }, |
| { |
| "epoch": 1.138095238095238, |
| "grad_norm": 0.3679749667644501, |
| "learning_rate": 7.770334928229667e-06, |
| "loss": 0.053099337965250015, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05453, |
| "step": 239, |
| "tokens/total": 978944, |
| "tokens/train_per_sec_per_gpu": 11.86, |
| "tokens/trainable": 70324 |
| }, |
| { |
| "epoch": 1.1428571428571428, |
| "grad_norm": 0.9312533140182495, |
| "learning_rate": 7.760765550239236e-06, |
| "loss": 0.16996754705905914, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.18527, |
| "step": 240, |
| "tokens/total": 983040, |
| "tokens/train_per_sec_per_gpu": 13.42, |
| "tokens/trainable": 70667 |
| }, |
| { |
| "epoch": 1.1476190476190475, |
| "grad_norm": 0.8272965550422668, |
| "learning_rate": 7.751196172248805e-06, |
| "loss": 0.1209377869963646, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12855, |
| "step": 241, |
| "tokens/total": 987136, |
| "tokens/train_per_sec_per_gpu": 12.75, |
| "tokens/trainable": 70927 |
| }, |
| { |
| "epoch": 1.1523809523809523, |
| "grad_norm": 1.1589566469192505, |
| "learning_rate": 7.741626794258373e-06, |
| "loss": 0.16512207686901093, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17954, |
| "step": 242, |
| "tokens/total": 991232, |
| "tokens/train_per_sec_per_gpu": 13.03, |
| "tokens/trainable": 71234 |
| }, |
| { |
| "epoch": 1.157142857142857, |
| "grad_norm": 0.7066253423690796, |
| "learning_rate": 7.732057416267943e-06, |
| "loss": 0.08280467987060547, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08633, |
| "step": 243, |
| "tokens/total": 995328, |
| "tokens/train_per_sec_per_gpu": 14.93, |
| "tokens/trainable": 71606 |
| }, |
| { |
| "epoch": 1.161904761904762, |
| "grad_norm": 0.4909917414188385, |
| "learning_rate": 7.722488038277512e-06, |
| "loss": 0.10505213588476181, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11077, |
| "step": 244, |
| "tokens/total": 999424, |
| "tokens/train_per_sec_per_gpu": 16.84, |
| "tokens/trainable": 71984 |
| }, |
| { |
| "epoch": 1.1666666666666667, |
| "grad_norm": 0.8122262954711914, |
| "learning_rate": 7.712918660287081e-06, |
| "loss": 0.23099766671657562, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.25986, |
| "step": 245, |
| "tokens/total": 1003520, |
| "tokens/train_per_sec_per_gpu": 20.06, |
| "tokens/trainable": 72353 |
| }, |
| { |
| "epoch": 1.1714285714285715, |
| "grad_norm": 0.44005194306373596, |
| "learning_rate": 7.703349282296652e-06, |
| "loss": 0.08119498938322067, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08458, |
| "step": 246, |
| "tokens/total": 1007616, |
| "tokens/train_per_sec_per_gpu": 17.01, |
| "tokens/trainable": 72628 |
| }, |
| { |
| "epoch": 1.1761904761904762, |
| "grad_norm": 0.672211229801178, |
| "learning_rate": 7.693779904306222e-06, |
| "loss": 0.14602981507778168, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15723, |
| "step": 247, |
| "tokens/total": 1011712, |
| "tokens/train_per_sec_per_gpu": 8.19, |
| "tokens/trainable": 72922 |
| }, |
| { |
| "epoch": 1.180952380952381, |
| "grad_norm": 0.734875500202179, |
| "learning_rate": 7.68421052631579e-06, |
| "loss": 0.13793110847473145, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1479, |
| "step": 248, |
| "tokens/total": 1015808, |
| "tokens/train_per_sec_per_gpu": 12.01, |
| "tokens/trainable": 73196 |
| }, |
| { |
| "epoch": 1.1857142857142857, |
| "grad_norm": 0.9989974498748779, |
| "learning_rate": 7.674641148325359e-06, |
| "loss": 0.10894070565700531, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1151, |
| "step": 249, |
| "tokens/total": 1019904, |
| "tokens/train_per_sec_per_gpu": 11.97, |
| "tokens/trainable": 73462 |
| }, |
| { |
| "epoch": 1.1904761904761905, |
| "grad_norm": 0.5371176600456238, |
| "learning_rate": 7.665071770334928e-06, |
| "loss": 0.09055919945240021, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09479, |
| "step": 250, |
| "tokens/total": 1024000, |
| "tokens/train_per_sec_per_gpu": 18.03, |
| "tokens/trainable": 73804 |
| }, |
| { |
| "epoch": 1.1952380952380952, |
| "grad_norm": 0.5338223576545715, |
| "learning_rate": 7.655502392344498e-06, |
| "loss": 0.1184496283531189, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12575, |
| "step": 251, |
| "tokens/total": 1028096, |
| "tokens/train_per_sec_per_gpu": 14.19, |
| "tokens/trainable": 74099 |
| }, |
| { |
| "epoch": 1.2, |
| "grad_norm": 0.7946963310241699, |
| "learning_rate": 7.645933014354067e-06, |
| "loss": 0.1663852483034134, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.18103, |
| "step": 252, |
| "tokens/total": 1032192, |
| "tokens/train_per_sec_per_gpu": 13.78, |
| "tokens/trainable": 74370 |
| }, |
| { |
| "epoch": 1.2047619047619047, |
| "grad_norm": 0.5551656484603882, |
| "learning_rate": 7.636363636363638e-06, |
| "loss": 0.07018721103668213, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07271, |
| "step": 253, |
| "tokens/total": 1036288, |
| "tokens/train_per_sec_per_gpu": 11.46, |
| "tokens/trainable": 74627 |
| }, |
| { |
| "epoch": 1.2095238095238094, |
| "grad_norm": 0.7267068028450012, |
| "learning_rate": 7.626794258373207e-06, |
| "loss": 0.14088734984397888, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15129, |
| "step": 254, |
| "tokens/total": 1040384, |
| "tokens/train_per_sec_per_gpu": 14.73, |
| "tokens/trainable": 74952 |
| }, |
| { |
| "epoch": 1.2142857142857142, |
| "grad_norm": 0.8808903694152832, |
| "learning_rate": 7.617224880382775e-06, |
| "loss": 0.0552162267267704, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05677, |
| "step": 255, |
| "tokens/total": 1044480, |
| "tokens/train_per_sec_per_gpu": 9.23, |
| "tokens/trainable": 75179 |
| }, |
| { |
| "epoch": 1.2190476190476192, |
| "grad_norm": 0.6379745006561279, |
| "learning_rate": 7.6076555023923446e-06, |
| "loss": 0.09947262704372406, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10459, |
| "step": 256, |
| "tokens/total": 1048576, |
| "tokens/train_per_sec_per_gpu": 10.05, |
| "tokens/trainable": 75427 |
| }, |
| { |
| "epoch": 1.223809523809524, |
| "grad_norm": 0.8014132976531982, |
| "learning_rate": 7.598086124401915e-06, |
| "loss": 0.07943955808877945, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08268, |
| "step": 257, |
| "tokens/total": 1052672, |
| "tokens/train_per_sec_per_gpu": 10.22, |
| "tokens/trainable": 75700 |
| }, |
| { |
| "epoch": 1.2285714285714286, |
| "grad_norm": 0.6632212996482849, |
| "learning_rate": 7.588516746411484e-06, |
| "loss": 0.1634090542793274, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17752, |
| "step": 258, |
| "tokens/total": 1056768, |
| "tokens/train_per_sec_per_gpu": 13.72, |
| "tokens/trainable": 76060 |
| }, |
| { |
| "epoch": 1.2333333333333334, |
| "grad_norm": 0.692261278629303, |
| "learning_rate": 7.578947368421054e-06, |
| "loss": 0.10278518497943878, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10825, |
| "step": 259, |
| "tokens/total": 1060864, |
| "tokens/train_per_sec_per_gpu": 12.97, |
| "tokens/trainable": 76309 |
| }, |
| { |
| "epoch": 1.2380952380952381, |
| "grad_norm": 0.9587100148200989, |
| "learning_rate": 7.569377990430623e-06, |
| "loss": 0.13651782274246216, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.14628, |
| "step": 260, |
| "tokens/total": 1064960, |
| "tokens/train_per_sec_per_gpu": 11.4, |
| "tokens/trainable": 76541 |
| }, |
| { |
| "epoch": 1.2428571428571429, |
| "grad_norm": 1.1532655954360962, |
| "learning_rate": 7.5598086124401915e-06, |
| "loss": 0.1658339500427246, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.18038, |
| "step": 261, |
| "tokens/total": 1069056, |
| "tokens/train_per_sec_per_gpu": 15.11, |
| "tokens/trainable": 76799 |
| }, |
| { |
| "epoch": 1.2476190476190476, |
| "grad_norm": 0.6354696750640869, |
| "learning_rate": 7.550239234449761e-06, |
| "loss": 0.09453555941581726, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09915, |
| "step": 262, |
| "tokens/total": 1073152, |
| "tokens/train_per_sec_per_gpu": 11.51, |
| "tokens/trainable": 77070 |
| }, |
| { |
| "epoch": 1.2523809523809524, |
| "grad_norm": 0.5260769724845886, |
| "learning_rate": 7.54066985645933e-06, |
| "loss": 0.060988280922174454, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.06289, |
| "step": 263, |
| "tokens/total": 1077248, |
| "tokens/train_per_sec_per_gpu": 11.93, |
| "tokens/trainable": 77323 |
| }, |
| { |
| "epoch": 1.2571428571428571, |
| "grad_norm": 0.7785736918449402, |
| "learning_rate": 7.5311004784689005e-06, |
| "loss": 0.10227103531360626, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10768, |
| "step": 264, |
| "tokens/total": 1081344, |
| "tokens/train_per_sec_per_gpu": 10.84, |
| "tokens/trainable": 77599 |
| }, |
| { |
| "epoch": 1.2619047619047619, |
| "grad_norm": 0.5048001408576965, |
| "learning_rate": 7.52153110047847e-06, |
| "loss": 0.065130315721035, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0673, |
| "step": 265, |
| "tokens/total": 1085440, |
| "tokens/train_per_sec_per_gpu": 10.88, |
| "tokens/trainable": 77842 |
| }, |
| { |
| "epoch": 1.2666666666666666, |
| "grad_norm": 0.6458641886711121, |
| "learning_rate": 7.511961722488039e-06, |
| "loss": 0.1348947137594223, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14442, |
| "step": 266, |
| "tokens/total": 1089536, |
| "tokens/train_per_sec_per_gpu": 17.6, |
| "tokens/trainable": 78180 |
| }, |
| { |
| "epoch": 1.2714285714285714, |
| "grad_norm": 0.6480576992034912, |
| "learning_rate": 7.502392344497609e-06, |
| "loss": 0.10622403025627136, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11207, |
| "step": 267, |
| "tokens/total": 1093632, |
| "tokens/train_per_sec_per_gpu": 13.68, |
| "tokens/trainable": 78442 |
| }, |
| { |
| "epoch": 1.276190476190476, |
| "grad_norm": 0.6444092392921448, |
| "learning_rate": 7.492822966507177e-06, |
| "loss": 0.09460228681564331, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09922, |
| "step": 268, |
| "tokens/total": 1097728, |
| "tokens/train_per_sec_per_gpu": 8.4, |
| "tokens/trainable": 78660 |
| }, |
| { |
| "epoch": 1.2809523809523808, |
| "grad_norm": 0.6892173886299133, |
| "learning_rate": 7.483253588516747e-06, |
| "loss": 0.13539853692054749, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14499, |
| "step": 269, |
| "tokens/total": 1101824, |
| "tokens/train_per_sec_per_gpu": 9.93, |
| "tokens/trainable": 78945 |
| }, |
| { |
| "epoch": 1.2857142857142856, |
| "grad_norm": 0.7260250449180603, |
| "learning_rate": 7.473684210526316e-06, |
| "loss": 0.09833601862192154, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10333, |
| "step": 270, |
| "tokens/total": 1105920, |
| "tokens/train_per_sec_per_gpu": 10.96, |
| "tokens/trainable": 79182 |
| }, |
| { |
| "epoch": 1.2904761904761906, |
| "grad_norm": 0.7841465473175049, |
| "learning_rate": 7.464114832535886e-06, |
| "loss": 0.11291639506816864, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11954, |
| "step": 271, |
| "tokens/total": 1110016, |
| "tokens/train_per_sec_per_gpu": 13.7, |
| "tokens/trainable": 79453 |
| }, |
| { |
| "epoch": 1.2952380952380953, |
| "grad_norm": 0.8338773250579834, |
| "learning_rate": 7.454545454545456e-06, |
| "loss": 0.18002504110336304, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.19725, |
| "step": 272, |
| "tokens/total": 1114112, |
| "tokens/train_per_sec_per_gpu": 16.17, |
| "tokens/trainable": 79767 |
| }, |
| { |
| "epoch": 1.3, |
| "grad_norm": 0.6367767453193665, |
| "learning_rate": 7.444976076555025e-06, |
| "loss": 0.09027963876724243, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09448, |
| "step": 273, |
| "tokens/total": 1118208, |
| "tokens/train_per_sec_per_gpu": 16.24, |
| "tokens/trainable": 80046 |
| }, |
| { |
| "epoch": 1.3047619047619048, |
| "grad_norm": 1.0161387920379639, |
| "learning_rate": 7.4354066985645936e-06, |
| "loss": 0.1279660016298294, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13651, |
| "step": 274, |
| "tokens/total": 1122304, |
| "tokens/train_per_sec_per_gpu": 10.88, |
| "tokens/trainable": 80319 |
| }, |
| { |
| "epoch": 1.3095238095238095, |
| "grad_norm": 1.0113588571548462, |
| "learning_rate": 7.425837320574163e-06, |
| "loss": 0.1560673713684082, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1689, |
| "step": 275, |
| "tokens/total": 1126400, |
| "tokens/train_per_sec_per_gpu": 12.7, |
| "tokens/trainable": 80627 |
| }, |
| { |
| "epoch": 1.3142857142857143, |
| "grad_norm": 0.6805703639984131, |
| "learning_rate": 7.416267942583732e-06, |
| "loss": 0.08896687626838684, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09304, |
| "step": 276, |
| "tokens/total": 1130496, |
| "tokens/train_per_sec_per_gpu": 12.44, |
| "tokens/trainable": 80874 |
| }, |
| { |
| "epoch": 1.319047619047619, |
| "grad_norm": 0.8909221291542053, |
| "learning_rate": 7.406698564593302e-06, |
| "loss": 0.17334583401679993, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.18928, |
| "step": 277, |
| "tokens/total": 1134592, |
| "tokens/train_per_sec_per_gpu": 15.86, |
| "tokens/trainable": 81176 |
| }, |
| { |
| "epoch": 1.3238095238095238, |
| "grad_norm": 0.8896336555480957, |
| "learning_rate": 7.397129186602872e-06, |
| "loss": 0.16712436079978943, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1819, |
| "step": 278, |
| "tokens/total": 1138688, |
| "tokens/train_per_sec_per_gpu": 12.49, |
| "tokens/trainable": 81474 |
| }, |
| { |
| "epoch": 1.3285714285714285, |
| "grad_norm": 0.764587938785553, |
| "learning_rate": 7.387559808612441e-06, |
| "loss": 0.12825949490070343, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13685, |
| "step": 279, |
| "tokens/total": 1142784, |
| "tokens/train_per_sec_per_gpu": 11.78, |
| "tokens/trainable": 81741 |
| }, |
| { |
| "epoch": 1.3333333333333333, |
| "grad_norm": 0.6739766001701355, |
| "learning_rate": 7.377990430622011e-06, |
| "loss": 0.1249503344297409, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13309, |
| "step": 280, |
| "tokens/total": 1146880, |
| "tokens/train_per_sec_per_gpu": 15.73, |
| "tokens/trainable": 82115 |
| }, |
| { |
| "epoch": 1.3380952380952382, |
| "grad_norm": 0.4460581839084625, |
| "learning_rate": 7.368421052631579e-06, |
| "loss": 0.07482986897230148, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0777, |
| "step": 281, |
| "tokens/total": 1150976, |
| "tokens/train_per_sec_per_gpu": 12.85, |
| "tokens/trainable": 82422 |
| }, |
| { |
| "epoch": 1.342857142857143, |
| "grad_norm": 0.8023619055747986, |
| "learning_rate": 7.358851674641149e-06, |
| "loss": 0.09925596415996552, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10435, |
| "step": 282, |
| "tokens/total": 1155072, |
| "tokens/train_per_sec_per_gpu": 14.95, |
| "tokens/trainable": 82709 |
| }, |
| { |
| "epoch": 1.3476190476190477, |
| "grad_norm": 0.6830031275749207, |
| "learning_rate": 7.349282296650718e-06, |
| "loss": 0.08342678844928741, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08701, |
| "step": 283, |
| "tokens/total": 1159168, |
| "tokens/train_per_sec_per_gpu": 13.49, |
| "tokens/trainable": 82977 |
| }, |
| { |
| "epoch": 1.3523809523809525, |
| "grad_norm": 0.8009830713272095, |
| "learning_rate": 7.3397129186602874e-06, |
| "loss": 0.07836952805519104, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08152, |
| "step": 284, |
| "tokens/total": 1163264, |
| "tokens/train_per_sec_per_gpu": 12.22, |
| "tokens/trainable": 83276 |
| }, |
| { |
| "epoch": 1.3571428571428572, |
| "grad_norm": 0.7553001642227173, |
| "learning_rate": 7.330143540669858e-06, |
| "loss": 0.1220192164182663, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12978, |
| "step": 285, |
| "tokens/total": 1167360, |
| "tokens/train_per_sec_per_gpu": 10.55, |
| "tokens/trainable": 83542 |
| }, |
| { |
| "epoch": 1.361904761904762, |
| "grad_norm": 0.44703495502471924, |
| "learning_rate": 7.320574162679427e-06, |
| "loss": 0.08532048761844635, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08907, |
| "step": 286, |
| "tokens/total": 1171456, |
| "tokens/train_per_sec_per_gpu": 13.86, |
| "tokens/trainable": 83900 |
| }, |
| { |
| "epoch": 1.3666666666666667, |
| "grad_norm": 1.0438194274902344, |
| "learning_rate": 7.311004784688996e-06, |
| "loss": 0.07847965508699417, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08164, |
| "step": 287, |
| "tokens/total": 1175552, |
| "tokens/train_per_sec_per_gpu": 16.2, |
| "tokens/trainable": 84167 |
| }, |
| { |
| "epoch": 1.3714285714285714, |
| "grad_norm": 0.7424266338348389, |
| "learning_rate": 7.301435406698565e-06, |
| "loss": 0.10971373319625854, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11596, |
| "step": 288, |
| "tokens/total": 1179648, |
| "tokens/train_per_sec_per_gpu": 14.85, |
| "tokens/trainable": 84457 |
| }, |
| { |
| "epoch": 1.3761904761904762, |
| "grad_norm": 0.38364148139953613, |
| "learning_rate": 7.291866028708134e-06, |
| "loss": 0.03297719731926918, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.03353, |
| "step": 289, |
| "tokens/total": 1183744, |
| "tokens/train_per_sec_per_gpu": 11.14, |
| "tokens/trainable": 84681 |
| }, |
| { |
| "epoch": 1.380952380952381, |
| "grad_norm": 0.8590928912162781, |
| "learning_rate": 7.282296650717704e-06, |
| "loss": 0.14657095074653625, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15786, |
| "step": 290, |
| "tokens/total": 1187840, |
| "tokens/train_per_sec_per_gpu": 15.12, |
| "tokens/trainable": 84973 |
| }, |
| { |
| "epoch": 1.3857142857142857, |
| "grad_norm": 0.3598606586456299, |
| "learning_rate": 7.272727272727273e-06, |
| "loss": 0.08855046331882477, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09259, |
| "step": 291, |
| "tokens/total": 1191936, |
| "tokens/train_per_sec_per_gpu": 22.23, |
| "tokens/trainable": 85410 |
| }, |
| { |
| "epoch": 1.3904761904761904, |
| "grad_norm": 0.5362772345542908, |
| "learning_rate": 7.263157894736843e-06, |
| "loss": 0.06545418500900269, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06764, |
| "step": 292, |
| "tokens/total": 1196032, |
| "tokens/train_per_sec_per_gpu": 14.16, |
| "tokens/trainable": 85650 |
| }, |
| { |
| "epoch": 1.3952380952380952, |
| "grad_norm": 0.8264137506484985, |
| "learning_rate": 7.253588516746413e-06, |
| "loss": 0.10715029388666153, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1131, |
| "step": 293, |
| "tokens/total": 1200128, |
| "tokens/train_per_sec_per_gpu": 12.71, |
| "tokens/trainable": 85895 |
| }, |
| { |
| "epoch": 1.4, |
| "grad_norm": 0.7366301417350769, |
| "learning_rate": 7.244019138755981e-06, |
| "loss": 0.10185449570417404, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10722, |
| "step": 294, |
| "tokens/total": 1204224, |
| "tokens/train_per_sec_per_gpu": 11.18, |
| "tokens/trainable": 86116 |
| }, |
| { |
| "epoch": 1.4047619047619047, |
| "grad_norm": 0.7494551539421082, |
| "learning_rate": 7.234449760765551e-06, |
| "loss": 0.09909769147634506, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10417, |
| "step": 295, |
| "tokens/total": 1208320, |
| "tokens/train_per_sec_per_gpu": 9.06, |
| "tokens/trainable": 86414 |
| }, |
| { |
| "epoch": 1.4095238095238094, |
| "grad_norm": 0.8216342926025391, |
| "learning_rate": 7.22488038277512e-06, |
| "loss": 0.13356107473373413, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.14289, |
| "step": 296, |
| "tokens/total": 1212416, |
| "tokens/train_per_sec_per_gpu": 17.71, |
| "tokens/trainable": 86729 |
| }, |
| { |
| "epoch": 1.4142857142857144, |
| "grad_norm": 0.6832106709480286, |
| "learning_rate": 7.2153110047846895e-06, |
| "loss": 0.07904009521007538, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08225, |
| "step": 297, |
| "tokens/total": 1216512, |
| "tokens/train_per_sec_per_gpu": 11.21, |
| "tokens/trainable": 87033 |
| }, |
| { |
| "epoch": 1.4190476190476191, |
| "grad_norm": 0.635546863079071, |
| "learning_rate": 7.205741626794259e-06, |
| "loss": 0.061884887516498566, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06384, |
| "step": 298, |
| "tokens/total": 1220608, |
| "tokens/train_per_sec_per_gpu": 11.1, |
| "tokens/trainable": 87281 |
| }, |
| { |
| "epoch": 1.4238095238095239, |
| "grad_norm": 1.098975419998169, |
| "learning_rate": 7.196172248803829e-06, |
| "loss": 0.1470986008644104, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15847, |
| "step": 299, |
| "tokens/total": 1224704, |
| "tokens/train_per_sec_per_gpu": 14.32, |
| "tokens/trainable": 87573 |
| }, |
| { |
| "epoch": 1.4285714285714286, |
| "grad_norm": 0.7100666761398315, |
| "learning_rate": 7.186602870813398e-06, |
| "loss": 0.08846946060657501, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0925, |
| "step": 300, |
| "tokens/total": 1228800, |
| "tokens/train_per_sec_per_gpu": 11.7, |
| "tokens/trainable": 87829 |
| }, |
| { |
| "epoch": 1.4333333333333333, |
| "grad_norm": 0.8944647908210754, |
| "learning_rate": 7.177033492822967e-06, |
| "loss": 0.08944831788539886, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09357, |
| "step": 301, |
| "tokens/total": 1232896, |
| "tokens/train_per_sec_per_gpu": 10.27, |
| "tokens/trainable": 88067 |
| }, |
| { |
| "epoch": 1.438095238095238, |
| "grad_norm": 0.6894711256027222, |
| "learning_rate": 7.1674641148325364e-06, |
| "loss": 0.10426825284957886, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1099, |
| "step": 302, |
| "tokens/total": 1236992, |
| "tokens/train_per_sec_per_gpu": 13.61, |
| "tokens/trainable": 88342 |
| }, |
| { |
| "epoch": 1.4428571428571428, |
| "grad_norm": 1.4308092594146729, |
| "learning_rate": 7.157894736842106e-06, |
| "loss": 0.09522052109241486, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0999, |
| "step": 303, |
| "tokens/total": 1241088, |
| "tokens/train_per_sec_per_gpu": 11.01, |
| "tokens/trainable": 88603 |
| }, |
| { |
| "epoch": 1.4476190476190476, |
| "grad_norm": 0.7011018395423889, |
| "learning_rate": 7.148325358851675e-06, |
| "loss": 0.06806151568889618, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07043, |
| "step": 304, |
| "tokens/total": 1245184, |
| "tokens/train_per_sec_per_gpu": 11.0, |
| "tokens/trainable": 88843 |
| }, |
| { |
| "epoch": 1.4523809523809523, |
| "grad_norm": 0.885724663734436, |
| "learning_rate": 7.138755980861245e-06, |
| "loss": 0.16685882210731506, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.18159, |
| "step": 305, |
| "tokens/total": 1249280, |
| "tokens/train_per_sec_per_gpu": 12.19, |
| "tokens/trainable": 89102 |
| }, |
| { |
| "epoch": 1.457142857142857, |
| "grad_norm": 0.4040443003177643, |
| "learning_rate": 7.129186602870813e-06, |
| "loss": 0.07094176113605499, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.07352, |
| "step": 306, |
| "tokens/total": 1253376, |
| "tokens/train_per_sec_per_gpu": 8.89, |
| "tokens/trainable": 89437 |
| }, |
| { |
| "epoch": 1.461904761904762, |
| "grad_norm": 0.8666446805000305, |
| "learning_rate": 7.119617224880383e-06, |
| "loss": 0.11042168736457825, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11675, |
| "step": 307, |
| "tokens/total": 1257472, |
| "tokens/train_per_sec_per_gpu": 13.37, |
| "tokens/trainable": 89673 |
| }, |
| { |
| "epoch": 1.4666666666666668, |
| "grad_norm": 0.6336524486541748, |
| "learning_rate": 7.110047846889953e-06, |
| "loss": 0.08677862584590912, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09066, |
| "step": 308, |
| "tokens/total": 1261568, |
| "tokens/train_per_sec_per_gpu": 10.42, |
| "tokens/trainable": 89950 |
| }, |
| { |
| "epoch": 1.4714285714285715, |
| "grad_norm": 0.7790036797523499, |
| "learning_rate": 7.100478468899522e-06, |
| "loss": 0.0888119488954544, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09288, |
| "step": 309, |
| "tokens/total": 1265664, |
| "tokens/train_per_sec_per_gpu": 11.92, |
| "tokens/trainable": 90153 |
| }, |
| { |
| "epoch": 1.4761904761904763, |
| "grad_norm": 0.7459496855735779, |
| "learning_rate": 7.0909090909090916e-06, |
| "loss": 0.08131097257137299, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08471, |
| "step": 310, |
| "tokens/total": 1269760, |
| "tokens/train_per_sec_per_gpu": 11.39, |
| "tokens/trainable": 90399 |
| }, |
| { |
| "epoch": 1.480952380952381, |
| "grad_norm": 0.46430832147598267, |
| "learning_rate": 7.081339712918661e-06, |
| "loss": 0.039737895131111145, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04054, |
| "step": 311, |
| "tokens/total": 1273856, |
| "tokens/train_per_sec_per_gpu": 8.41, |
| "tokens/trainable": 90616 |
| }, |
| { |
| "epoch": 1.4857142857142858, |
| "grad_norm": 0.8076871037483215, |
| "learning_rate": 7.07177033492823e-06, |
| "loss": 0.09725730866193771, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10214, |
| "step": 312, |
| "tokens/total": 1277952, |
| "tokens/train_per_sec_per_gpu": 12.34, |
| "tokens/trainable": 90842 |
| }, |
| { |
| "epoch": 1.4904761904761905, |
| "grad_norm": 0.8504095077514648, |
| "learning_rate": 7.062200956937799e-06, |
| "loss": 0.13007959723472595, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13892, |
| "step": 313, |
| "tokens/total": 1282048, |
| "tokens/train_per_sec_per_gpu": 12.25, |
| "tokens/trainable": 91094 |
| }, |
| { |
| "epoch": 1.4952380952380953, |
| "grad_norm": 0.8776680827140808, |
| "learning_rate": 7.052631578947369e-06, |
| "loss": 0.09825856983661652, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10325, |
| "step": 314, |
| "tokens/total": 1286144, |
| "tokens/train_per_sec_per_gpu": 15.21, |
| "tokens/trainable": 91382 |
| }, |
| { |
| "epoch": 1.5, |
| "grad_norm": 0.540379524230957, |
| "learning_rate": 7.0430622009569385e-06, |
| "loss": 0.1302526891231537, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13912, |
| "step": 315, |
| "tokens/total": 1290240, |
| "tokens/train_per_sec_per_gpu": 21.51, |
| "tokens/trainable": 91745 |
| }, |
| { |
| "epoch": 1.5047619047619047, |
| "grad_norm": 1.0059609413146973, |
| "learning_rate": 7.033492822966508e-06, |
| "loss": 0.13155148923397064, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1406, |
| "step": 316, |
| "tokens/total": 1294336, |
| "tokens/train_per_sec_per_gpu": 12.26, |
| "tokens/trainable": 92034 |
| }, |
| { |
| "epoch": 1.5095238095238095, |
| "grad_norm": 0.8393654227256775, |
| "learning_rate": 7.023923444976077e-06, |
| "loss": 0.08929277956485748, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0934, |
| "step": 317, |
| "tokens/total": 1298432, |
| "tokens/train_per_sec_per_gpu": 9.35, |
| "tokens/trainable": 92367 |
| }, |
| { |
| "epoch": 1.5142857142857142, |
| "grad_norm": 0.48736438155174255, |
| "learning_rate": 7.014354066985647e-06, |
| "loss": 0.044491663575172424, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0455, |
| "step": 318, |
| "tokens/total": 1302528, |
| "tokens/train_per_sec_per_gpu": 13.14, |
| "tokens/trainable": 92640 |
| }, |
| { |
| "epoch": 1.519047619047619, |
| "grad_norm": 1.1108299493789673, |
| "learning_rate": 7.004784688995215e-06, |
| "loss": 0.15828374028205872, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1715, |
| "step": 319, |
| "tokens/total": 1306624, |
| "tokens/train_per_sec_per_gpu": 12.61, |
| "tokens/trainable": 92922 |
| }, |
| { |
| "epoch": 1.5238095238095237, |
| "grad_norm": 0.5160102248191833, |
| "learning_rate": 6.995215311004785e-06, |
| "loss": 0.0708310604095459, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0734, |
| "step": 320, |
| "tokens/total": 1310720, |
| "tokens/train_per_sec_per_gpu": 13.54, |
| "tokens/trainable": 93214 |
| }, |
| { |
| "epoch": 1.5285714285714285, |
| "grad_norm": 0.43335914611816406, |
| "learning_rate": 6.985645933014355e-06, |
| "loss": 0.055120449513196945, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05667, |
| "step": 321, |
| "tokens/total": 1314816, |
| "tokens/train_per_sec_per_gpu": 9.65, |
| "tokens/trainable": 93542 |
| }, |
| { |
| "epoch": 1.5333333333333332, |
| "grad_norm": 0.6859294176101685, |
| "learning_rate": 6.976076555023924e-06, |
| "loss": 0.09708588570356369, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10196, |
| "step": 322, |
| "tokens/total": 1318912, |
| "tokens/train_per_sec_per_gpu": 13.52, |
| "tokens/trainable": 93821 |
| }, |
| { |
| "epoch": 1.538095238095238, |
| "grad_norm": 1.008861780166626, |
| "learning_rate": 6.966507177033494e-06, |
| "loss": 0.1100694090127945, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11636, |
| "step": 323, |
| "tokens/total": 1323008, |
| "tokens/train_per_sec_per_gpu": 11.45, |
| "tokens/trainable": 94043 |
| }, |
| { |
| "epoch": 1.5428571428571427, |
| "grad_norm": 0.6240075826644897, |
| "learning_rate": 6.956937799043063e-06, |
| "loss": 0.09549560397863388, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1002, |
| "step": 324, |
| "tokens/total": 1327104, |
| "tokens/train_per_sec_per_gpu": 14.18, |
| "tokens/trainable": 94337 |
| }, |
| { |
| "epoch": 1.5476190476190477, |
| "grad_norm": 0.5021734833717346, |
| "learning_rate": 6.947368421052632e-06, |
| "loss": 0.05273163318634033, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05415, |
| "step": 325, |
| "tokens/total": 1331200, |
| "tokens/train_per_sec_per_gpu": 13.13, |
| "tokens/trainable": 94588 |
| }, |
| { |
| "epoch": 1.5523809523809524, |
| "grad_norm": 0.7321131229400635, |
| "learning_rate": 6.937799043062201e-06, |
| "loss": 0.1110309511423111, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11743, |
| "step": 326, |
| "tokens/total": 1335296, |
| "tokens/train_per_sec_per_gpu": 12.4, |
| "tokens/trainable": 94876 |
| }, |
| { |
| "epoch": 1.5571428571428572, |
| "grad_norm": 1.0234777927398682, |
| "learning_rate": 6.92822966507177e-06, |
| "loss": 0.13599176704883575, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14567, |
| "step": 327, |
| "tokens/total": 1339392, |
| "tokens/train_per_sec_per_gpu": 14.3, |
| "tokens/trainable": 95143 |
| }, |
| { |
| "epoch": 1.561904761904762, |
| "grad_norm": 0.6523513793945312, |
| "learning_rate": 6.9186602870813405e-06, |
| "loss": 0.0951455608010292, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09982, |
| "step": 328, |
| "tokens/total": 1343488, |
| "tokens/train_per_sec_per_gpu": 14.24, |
| "tokens/trainable": 95460 |
| }, |
| { |
| "epoch": 1.5666666666666667, |
| "grad_norm": 0.8506219387054443, |
| "learning_rate": 6.90909090909091e-06, |
| "loss": 0.11431875824928284, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12111, |
| "step": 329, |
| "tokens/total": 1347584, |
| "tokens/train_per_sec_per_gpu": 18.78, |
| "tokens/trainable": 95830 |
| }, |
| { |
| "epoch": 1.5714285714285714, |
| "grad_norm": 0.8775290846824646, |
| "learning_rate": 6.899521531100479e-06, |
| "loss": 0.09016980230808258, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09436, |
| "step": 330, |
| "tokens/total": 1351680, |
| "tokens/train_per_sec_per_gpu": 9.96, |
| "tokens/trainable": 96026 |
| }, |
| { |
| "epoch": 1.5761904761904761, |
| "grad_norm": 1.2778103351593018, |
| "learning_rate": 6.889952153110049e-06, |
| "loss": 0.19941449165344238, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.22069, |
| "step": 331, |
| "tokens/total": 1355776, |
| "tokens/train_per_sec_per_gpu": 17.3, |
| "tokens/trainable": 96495 |
| }, |
| { |
| "epoch": 1.580952380952381, |
| "grad_norm": 0.5732797980308533, |
| "learning_rate": 6.880382775119617e-06, |
| "loss": 0.057705655694007874, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0594, |
| "step": 332, |
| "tokens/total": 1359872, |
| "tokens/train_per_sec_per_gpu": 10.48, |
| "tokens/trainable": 96764 |
| }, |
| { |
| "epoch": 1.5857142857142859, |
| "grad_norm": 0.6136135458946228, |
| "learning_rate": 6.870813397129187e-06, |
| "loss": 0.07990540564060211, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08318, |
| "step": 333, |
| "tokens/total": 1363968, |
| "tokens/train_per_sec_per_gpu": 9.73, |
| "tokens/trainable": 97021 |
| }, |
| { |
| "epoch": 1.5904761904761906, |
| "grad_norm": 0.760566771030426, |
| "learning_rate": 6.861244019138756e-06, |
| "loss": 0.10733136534690857, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1133, |
| "step": 334, |
| "tokens/total": 1368064, |
| "tokens/train_per_sec_per_gpu": 11.17, |
| "tokens/trainable": 97274 |
| }, |
| { |
| "epoch": 1.5952380952380953, |
| "grad_norm": 0.7491924166679382, |
| "learning_rate": 6.851674641148326e-06, |
| "loss": 0.08976206183433533, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09391, |
| "step": 335, |
| "tokens/total": 1372160, |
| "tokens/train_per_sec_per_gpu": 13.43, |
| "tokens/trainable": 97552 |
| }, |
| { |
| "epoch": 1.6, |
| "grad_norm": 0.8733037114143372, |
| "learning_rate": 6.842105263157896e-06, |
| "loss": 0.09935858845710754, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10446, |
| "step": 336, |
| "tokens/total": 1376256, |
| "tokens/train_per_sec_per_gpu": 15.48, |
| "tokens/trainable": 97852 |
| }, |
| { |
| "epoch": 1.6047619047619048, |
| "grad_norm": 0.6176996827125549, |
| "learning_rate": 6.832535885167465e-06, |
| "loss": 0.1155165433883667, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12245, |
| "step": 337, |
| "tokens/total": 1380352, |
| "tokens/train_per_sec_per_gpu": 8.8, |
| "tokens/trainable": 98089 |
| }, |
| { |
| "epoch": 1.6095238095238096, |
| "grad_norm": 0.556769073009491, |
| "learning_rate": 6.8229665071770344e-06, |
| "loss": 0.06867925822734833, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07109, |
| "step": 338, |
| "tokens/total": 1384448, |
| "tokens/train_per_sec_per_gpu": 10.65, |
| "tokens/trainable": 98331 |
| }, |
| { |
| "epoch": 1.6142857142857143, |
| "grad_norm": 1.0706437826156616, |
| "learning_rate": 6.813397129186603e-06, |
| "loss": 0.11363160610198975, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12034, |
| "step": 339, |
| "tokens/total": 1388544, |
| "tokens/train_per_sec_per_gpu": 20.54, |
| "tokens/trainable": 98733 |
| }, |
| { |
| "epoch": 1.619047619047619, |
| "grad_norm": 1.0246440172195435, |
| "learning_rate": 6.803827751196172e-06, |
| "loss": 0.09106756746768951, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09534, |
| "step": 340, |
| "tokens/total": 1392640, |
| "tokens/train_per_sec_per_gpu": 9.49, |
| "tokens/trainable": 98929 |
| }, |
| { |
| "epoch": 1.6238095238095238, |
| "grad_norm": 0.8117942810058594, |
| "learning_rate": 6.794258373205742e-06, |
| "loss": 0.08481201529502869, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08851, |
| "step": 341, |
| "tokens/total": 1396736, |
| "tokens/train_per_sec_per_gpu": 10.45, |
| "tokens/trainable": 99202 |
| }, |
| { |
| "epoch": 1.6285714285714286, |
| "grad_norm": 0.44152283668518066, |
| "learning_rate": 6.784688995215312e-06, |
| "loss": 0.06555698066949844, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06775, |
| "step": 342, |
| "tokens/total": 1400832, |
| "tokens/train_per_sec_per_gpu": 14.83, |
| "tokens/trainable": 99485 |
| }, |
| { |
| "epoch": 1.6333333333333333, |
| "grad_norm": 0.9880382418632507, |
| "learning_rate": 6.775119617224881e-06, |
| "loss": 0.1196918860077858, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12715, |
| "step": 343, |
| "tokens/total": 1404928, |
| "tokens/train_per_sec_per_gpu": 12.95, |
| "tokens/trainable": 99779 |
| }, |
| { |
| "epoch": 1.638095238095238, |
| "grad_norm": 1.3182851076126099, |
| "learning_rate": 6.765550239234451e-06, |
| "loss": 0.17728012800216675, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.19397, |
| "step": 344, |
| "tokens/total": 1409024, |
| "tokens/train_per_sec_per_gpu": 10.02, |
| "tokens/trainable": 100081 |
| }, |
| { |
| "epoch": 1.6428571428571428, |
| "grad_norm": 0.9957773089408875, |
| "learning_rate": 6.755980861244019e-06, |
| "loss": 0.15881581604480743, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17212, |
| "step": 345, |
| "tokens/total": 1413120, |
| "tokens/train_per_sec_per_gpu": 9.73, |
| "tokens/trainable": 100286 |
| }, |
| { |
| "epoch": 1.6476190476190475, |
| "grad_norm": 0.7199157476425171, |
| "learning_rate": 6.746411483253589e-06, |
| "loss": 0.10562394559383392, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1114, |
| "step": 346, |
| "tokens/total": 1417216, |
| "tokens/train_per_sec_per_gpu": 19.72, |
| "tokens/trainable": 100673 |
| }, |
| { |
| "epoch": 1.6523809523809523, |
| "grad_norm": 0.4368532598018646, |
| "learning_rate": 6.736842105263158e-06, |
| "loss": 0.0761297196149826, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0791, |
| "step": 347, |
| "tokens/total": 1421312, |
| "tokens/train_per_sec_per_gpu": 20.81, |
| "tokens/trainable": 101035 |
| }, |
| { |
| "epoch": 1.657142857142857, |
| "grad_norm": 0.8100814819335938, |
| "learning_rate": 6.7272727272727275e-06, |
| "loss": 0.1824449598789215, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.20015, |
| "step": 348, |
| "tokens/total": 1425408, |
| "tokens/train_per_sec_per_gpu": 14.81, |
| "tokens/trainable": 101374 |
| }, |
| { |
| "epoch": 1.6619047619047618, |
| "grad_norm": 0.9584652781486511, |
| "learning_rate": 6.717703349282298e-06, |
| "loss": 0.12359898537397385, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.13156, |
| "step": 349, |
| "tokens/total": 1429504, |
| "tokens/train_per_sec_per_gpu": 21.06, |
| "tokens/trainable": 101684 |
| }, |
| { |
| "epoch": 1.6666666666666665, |
| "grad_norm": 0.3303496539592743, |
| "learning_rate": 6.708133971291867e-06, |
| "loss": 0.06981052458286285, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0723, |
| "step": 350, |
| "tokens/total": 1433600, |
| "tokens/train_per_sec_per_gpu": 16.0, |
| "tokens/trainable": 102057 |
| }, |
| { |
| "epoch": 1.6714285714285713, |
| "grad_norm": 0.8605956435203552, |
| "learning_rate": 6.6985645933014365e-06, |
| "loss": 0.118899405002594, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12626, |
| "step": 351, |
| "tokens/total": 1437696, |
| "tokens/train_per_sec_per_gpu": 12.13, |
| "tokens/trainable": 102350 |
| }, |
| { |
| "epoch": 1.6761904761904762, |
| "grad_norm": 1.0154688358306885, |
| "learning_rate": 6.688995215311005e-06, |
| "loss": 0.11813990771770477, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1254, |
| "step": 352, |
| "tokens/total": 1441792, |
| "tokens/train_per_sec_per_gpu": 12.92, |
| "tokens/trainable": 102675 |
| }, |
| { |
| "epoch": 1.680952380952381, |
| "grad_norm": 0.8970211744308472, |
| "learning_rate": 6.679425837320574e-06, |
| "loss": 0.14539587497711182, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1565, |
| "step": 353, |
| "tokens/total": 1445888, |
| "tokens/train_per_sec_per_gpu": 13.5, |
| "tokens/trainable": 102963 |
| }, |
| { |
| "epoch": 1.6857142857142857, |
| "grad_norm": 0.8070185780525208, |
| "learning_rate": 6.669856459330144e-06, |
| "loss": 0.09510502219200134, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09977, |
| "step": 354, |
| "tokens/total": 1449984, |
| "tokens/train_per_sec_per_gpu": 9.91, |
| "tokens/trainable": 103183 |
| }, |
| { |
| "epoch": 1.6904761904761905, |
| "grad_norm": 0.7684424519538879, |
| "learning_rate": 6.660287081339713e-06, |
| "loss": 0.07780720293521881, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08091, |
| "step": 355, |
| "tokens/total": 1454080, |
| "tokens/train_per_sec_per_gpu": 10.73, |
| "tokens/trainable": 103388 |
| }, |
| { |
| "epoch": 1.6952380952380952, |
| "grad_norm": 0.4272719919681549, |
| "learning_rate": 6.6507177033492834e-06, |
| "loss": 0.05442621931433678, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05593, |
| "step": 356, |
| "tokens/total": 1458176, |
| "tokens/train_per_sec_per_gpu": 13.75, |
| "tokens/trainable": 103717 |
| }, |
| { |
| "epoch": 1.7, |
| "grad_norm": 0.6624967455863953, |
| "learning_rate": 6.641148325358853e-06, |
| "loss": 0.11500588059425354, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12188, |
| "step": 357, |
| "tokens/total": 1462272, |
| "tokens/train_per_sec_per_gpu": 11.4, |
| "tokens/trainable": 103954 |
| }, |
| { |
| "epoch": 1.704761904761905, |
| "grad_norm": 1.1226800680160522, |
| "learning_rate": 6.631578947368421e-06, |
| "loss": 0.13368532061576843, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.14303, |
| "step": 358, |
| "tokens/total": 1466368, |
| "tokens/train_per_sec_per_gpu": 11.12, |
| "tokens/trainable": 104220 |
| }, |
| { |
| "epoch": 1.7095238095238097, |
| "grad_norm": 1.0672284364700317, |
| "learning_rate": 6.622009569377991e-06, |
| "loss": 0.12235910445451736, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13016, |
| "step": 359, |
| "tokens/total": 1470464, |
| "tokens/train_per_sec_per_gpu": 19.02, |
| "tokens/trainable": 104555 |
| }, |
| { |
| "epoch": 1.7142857142857144, |
| "grad_norm": 0.9157024621963501, |
| "learning_rate": 6.61244019138756e-06, |
| "loss": 0.10388430953025818, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10947, |
| "step": 360, |
| "tokens/total": 1474560, |
| "tokens/train_per_sec_per_gpu": 18.68, |
| "tokens/trainable": 104964 |
| }, |
| { |
| "epoch": 1.7190476190476192, |
| "grad_norm": 1.042858600616455, |
| "learning_rate": 6.6028708133971295e-06, |
| "loss": 0.13133081793785095, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.14034, |
| "step": 361, |
| "tokens/total": 1478656, |
| "tokens/train_per_sec_per_gpu": 12.51, |
| "tokens/trainable": 105246 |
| }, |
| { |
| "epoch": 1.723809523809524, |
| "grad_norm": 0.787800669670105, |
| "learning_rate": 6.593301435406699e-06, |
| "loss": 0.11535128951072693, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12227, |
| "step": 362, |
| "tokens/total": 1482752, |
| "tokens/train_per_sec_per_gpu": 13.81, |
| "tokens/trainable": 105524 |
| }, |
| { |
| "epoch": 1.7285714285714286, |
| "grad_norm": 0.6248174905776978, |
| "learning_rate": 6.583732057416269e-06, |
| "loss": 0.08816708624362946, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09217, |
| "step": 363, |
| "tokens/total": 1486848, |
| "tokens/train_per_sec_per_gpu": 14.94, |
| "tokens/trainable": 105830 |
| }, |
| { |
| "epoch": 1.7333333333333334, |
| "grad_norm": 0.7269876599311829, |
| "learning_rate": 6.5741626794258385e-06, |
| "loss": 0.1025102511048317, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10795, |
| "step": 364, |
| "tokens/total": 1490944, |
| "tokens/train_per_sec_per_gpu": 14.06, |
| "tokens/trainable": 106111 |
| }, |
| { |
| "epoch": 1.7380952380952381, |
| "grad_norm": 0.8046413064002991, |
| "learning_rate": 6.564593301435407e-06, |
| "loss": 0.07084720581769943, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07342, |
| "step": 365, |
| "tokens/total": 1495040, |
| "tokens/train_per_sec_per_gpu": 14.63, |
| "tokens/trainable": 106357 |
| }, |
| { |
| "epoch": 1.7428571428571429, |
| "grad_norm": 0.7060100436210632, |
| "learning_rate": 6.5550239234449765e-06, |
| "loss": 0.13040365278720856, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13929, |
| "step": 366, |
| "tokens/total": 1499136, |
| "tokens/train_per_sec_per_gpu": 10.8, |
| "tokens/trainable": 106618 |
| }, |
| { |
| "epoch": 1.7476190476190476, |
| "grad_norm": 0.5565487146377563, |
| "learning_rate": 6.545454545454546e-06, |
| "loss": 0.06260109692811966, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.0646, |
| "step": 367, |
| "tokens/total": 1503232, |
| "tokens/train_per_sec_per_gpu": 9.03, |
| "tokens/trainable": 106851 |
| }, |
| { |
| "epoch": 1.7523809523809524, |
| "grad_norm": 0.7016215920448303, |
| "learning_rate": 6.535885167464115e-06, |
| "loss": 0.13496726751327515, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1445, |
| "step": 368, |
| "tokens/total": 1507328, |
| "tokens/train_per_sec_per_gpu": 14.23, |
| "tokens/trainable": 107133 |
| }, |
| { |
| "epoch": 1.7571428571428571, |
| "grad_norm": 0.9895106554031372, |
| "learning_rate": 6.526315789473685e-06, |
| "loss": 0.0984669178724289, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10348, |
| "step": 369, |
| "tokens/total": 1511424, |
| "tokens/train_per_sec_per_gpu": 15.05, |
| "tokens/trainable": 107493 |
| }, |
| { |
| "epoch": 1.7619047619047619, |
| "grad_norm": 0.8994340300559998, |
| "learning_rate": 6.516746411483255e-06, |
| "loss": 0.06861107051372528, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07102, |
| "step": 370, |
| "tokens/total": 1515520, |
| "tokens/train_per_sec_per_gpu": 18.0, |
| "tokens/trainable": 107792 |
| }, |
| { |
| "epoch": 1.7666666666666666, |
| "grad_norm": 0.8121207356452942, |
| "learning_rate": 6.507177033492823e-06, |
| "loss": 0.10615593940019608, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.112, |
| "step": 371, |
| "tokens/total": 1519616, |
| "tokens/train_per_sec_per_gpu": 14.3, |
| "tokens/trainable": 108092 |
| }, |
| { |
| "epoch": 1.7714285714285714, |
| "grad_norm": 0.7587819695472717, |
| "learning_rate": 6.497607655502393e-06, |
| "loss": 0.10872511565685272, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11486, |
| "step": 372, |
| "tokens/total": 1523712, |
| "tokens/train_per_sec_per_gpu": 19.98, |
| "tokens/trainable": 108466 |
| }, |
| { |
| "epoch": 1.776190476190476, |
| "grad_norm": 0.611366868019104, |
| "learning_rate": 6.488038277511962e-06, |
| "loss": 0.09398886561393738, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09855, |
| "step": 373, |
| "tokens/total": 1527808, |
| "tokens/train_per_sec_per_gpu": 15.17, |
| "tokens/trainable": 108833 |
| }, |
| { |
| "epoch": 1.7809523809523808, |
| "grad_norm": 0.6470115184783936, |
| "learning_rate": 6.4784688995215316e-06, |
| "loss": 0.07465333491563797, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07751, |
| "step": 374, |
| "tokens/total": 1531904, |
| "tokens/train_per_sec_per_gpu": 8.68, |
| "tokens/trainable": 109015 |
| }, |
| { |
| "epoch": 1.7857142857142856, |
| "grad_norm": 0.39367052912712097, |
| "learning_rate": 6.468899521531101e-06, |
| "loss": 0.07124529778957367, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07384, |
| "step": 375, |
| "tokens/total": 1536000, |
| "tokens/train_per_sec_per_gpu": 21.03, |
| "tokens/trainable": 109419 |
| }, |
| { |
| "epoch": 1.7904761904761903, |
| "grad_norm": 1.1016931533813477, |
| "learning_rate": 6.459330143540671e-06, |
| "loss": 0.11934101581573486, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12675, |
| "step": 376, |
| "tokens/total": 1540096, |
| "tokens/train_per_sec_per_gpu": 10.05, |
| "tokens/trainable": 109649 |
| }, |
| { |
| "epoch": 1.795238095238095, |
| "grad_norm": 0.6353244185447693, |
| "learning_rate": 6.449760765550239e-06, |
| "loss": 0.11098512262105942, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11738, |
| "step": 377, |
| "tokens/total": 1544192, |
| "tokens/train_per_sec_per_gpu": 13.61, |
| "tokens/trainable": 110032 |
| }, |
| { |
| "epoch": 1.8, |
| "grad_norm": 0.7181692719459534, |
| "learning_rate": 6.440191387559809e-06, |
| "loss": 0.10302020609378815, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10851, |
| "step": 378, |
| "tokens/total": 1548288, |
| "tokens/train_per_sec_per_gpu": 16.53, |
| "tokens/trainable": 110332 |
| }, |
| { |
| "epoch": 1.8047619047619048, |
| "grad_norm": 0.7024384140968323, |
| "learning_rate": 6.4306220095693785e-06, |
| "loss": 0.06668931245803833, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06896, |
| "step": 379, |
| "tokens/total": 1552384, |
| "tokens/train_per_sec_per_gpu": 10.82, |
| "tokens/trainable": 110580 |
| }, |
| { |
| "epoch": 1.8095238095238095, |
| "grad_norm": 0.6431664228439331, |
| "learning_rate": 6.421052631578948e-06, |
| "loss": 0.06305551528930664, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06509, |
| "step": 380, |
| "tokens/total": 1556480, |
| "tokens/train_per_sec_per_gpu": 14.2, |
| "tokens/trainable": 110882 |
| }, |
| { |
| "epoch": 1.8142857142857143, |
| "grad_norm": 0.7987076044082642, |
| "learning_rate": 6.411483253588517e-06, |
| "loss": 0.10767853260040283, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11369, |
| "step": 381, |
| "tokens/total": 1560576, |
| "tokens/train_per_sec_per_gpu": 13.11, |
| "tokens/trainable": 111160 |
| }, |
| { |
| "epoch": 1.819047619047619, |
| "grad_norm": 0.45520323514938354, |
| "learning_rate": 6.401913875598087e-06, |
| "loss": 0.03689165040850639, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03758, |
| "step": 382, |
| "tokens/total": 1564672, |
| "tokens/train_per_sec_per_gpu": 9.63, |
| "tokens/trainable": 111407 |
| }, |
| { |
| "epoch": 1.8238095238095238, |
| "grad_norm": 0.7428086996078491, |
| "learning_rate": 6.392344497607657e-06, |
| "loss": 0.09243252128362656, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09684, |
| "step": 383, |
| "tokens/total": 1568768, |
| "tokens/train_per_sec_per_gpu": 15.2, |
| "tokens/trainable": 111709 |
| }, |
| { |
| "epoch": 1.8285714285714287, |
| "grad_norm": 0.5691149234771729, |
| "learning_rate": 6.382775119617225e-06, |
| "loss": 0.09577442705631256, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10051, |
| "step": 384, |
| "tokens/total": 1572864, |
| "tokens/train_per_sec_per_gpu": 13.88, |
| "tokens/trainable": 112046 |
| }, |
| { |
| "epoch": 1.8333333333333335, |
| "grad_norm": 1.0002623796463013, |
| "learning_rate": 6.373205741626795e-06, |
| "loss": 0.10170154273509979, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10705, |
| "step": 385, |
| "tokens/total": 1576960, |
| "tokens/train_per_sec_per_gpu": 10.74, |
| "tokens/trainable": 112283 |
| }, |
| { |
| "epoch": 1.8380952380952382, |
| "grad_norm": 0.5939436554908752, |
| "learning_rate": 6.363636363636364e-06, |
| "loss": 0.06705500185489655, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06935, |
| "step": 386, |
| "tokens/total": 1581056, |
| "tokens/train_per_sec_per_gpu": 17.99, |
| "tokens/trainable": 112720 |
| }, |
| { |
| "epoch": 1.842857142857143, |
| "grad_norm": 0.6060131192207336, |
| "learning_rate": 6.354066985645934e-06, |
| "loss": 0.09325188398361206, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.09774, |
| "step": 387, |
| "tokens/total": 1585152, |
| "tokens/train_per_sec_per_gpu": 12.74, |
| "tokens/trainable": 113031 |
| }, |
| { |
| "epoch": 1.8476190476190477, |
| "grad_norm": 1.1861056089401245, |
| "learning_rate": 6.344497607655503e-06, |
| "loss": 0.13157960772514343, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14063, |
| "step": 388, |
| "tokens/total": 1589248, |
| "tokens/train_per_sec_per_gpu": 12.03, |
| "tokens/trainable": 113300 |
| }, |
| { |
| "epoch": 1.8523809523809525, |
| "grad_norm": 1.2170158624649048, |
| "learning_rate": 6.334928229665072e-06, |
| "loss": 0.14087001979351044, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15127, |
| "step": 389, |
| "tokens/total": 1593344, |
| "tokens/train_per_sec_per_gpu": 14.27, |
| "tokens/trainable": 113573 |
| }, |
| { |
| "epoch": 1.8571428571428572, |
| "grad_norm": 0.6340734362602234, |
| "learning_rate": 6.325358851674641e-06, |
| "loss": 0.11334896832704544, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12002, |
| "step": 390, |
| "tokens/total": 1597440, |
| "tokens/train_per_sec_per_gpu": 18.65, |
| "tokens/trainable": 113972 |
| }, |
| { |
| "epoch": 1.861904761904762, |
| "grad_norm": 0.6850183010101318, |
| "learning_rate": 6.31578947368421e-06, |
| "loss": 0.09625604748725891, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10104, |
| "step": 391, |
| "tokens/total": 1601536, |
| "tokens/train_per_sec_per_gpu": 11.8, |
| "tokens/trainable": 114303 |
| }, |
| { |
| "epoch": 1.8666666666666667, |
| "grad_norm": 0.7162591814994812, |
| "learning_rate": 6.3062200956937806e-06, |
| "loss": 0.09998545050621033, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10515, |
| "step": 392, |
| "tokens/total": 1605632, |
| "tokens/train_per_sec_per_gpu": 15.22, |
| "tokens/trainable": 114644 |
| }, |
| { |
| "epoch": 1.8714285714285714, |
| "grad_norm": 0.5838465690612793, |
| "learning_rate": 6.29665071770335e-06, |
| "loss": 0.06783585250377655, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.07019, |
| "step": 393, |
| "tokens/total": 1609728, |
| "tokens/train_per_sec_per_gpu": 15.64, |
| "tokens/trainable": 114985 |
| }, |
| { |
| "epoch": 1.8761904761904762, |
| "grad_norm": 0.8972572684288025, |
| "learning_rate": 6.287081339712919e-06, |
| "loss": 0.07981091737747192, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08308, |
| "step": 394, |
| "tokens/total": 1613824, |
| "tokens/train_per_sec_per_gpu": 10.59, |
| "tokens/trainable": 115295 |
| }, |
| { |
| "epoch": 1.880952380952381, |
| "grad_norm": 0.7543642520904541, |
| "learning_rate": 6.277511961722489e-06, |
| "loss": 0.11349859088659286, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12019, |
| "step": 395, |
| "tokens/total": 1617920, |
| "tokens/train_per_sec_per_gpu": 11.05, |
| "tokens/trainable": 115552 |
| }, |
| { |
| "epoch": 1.8857142857142857, |
| "grad_norm": 0.57085120677948, |
| "learning_rate": 6.267942583732058e-06, |
| "loss": 0.08496535569429398, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08868, |
| "step": 396, |
| "tokens/total": 1622016, |
| "tokens/train_per_sec_per_gpu": 10.2, |
| "tokens/trainable": 115799 |
| }, |
| { |
| "epoch": 1.8904761904761904, |
| "grad_norm": 0.6159265637397766, |
| "learning_rate": 6.258373205741627e-06, |
| "loss": 0.06171874701976776, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06366, |
| "step": 397, |
| "tokens/total": 1626112, |
| "tokens/train_per_sec_per_gpu": 12.61, |
| "tokens/trainable": 116117 |
| }, |
| { |
| "epoch": 1.8952380952380952, |
| "grad_norm": 0.8421717882156372, |
| "learning_rate": 6.248803827751196e-06, |
| "loss": 0.10187970101833344, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10725, |
| "step": 398, |
| "tokens/total": 1630208, |
| "tokens/train_per_sec_per_gpu": 15.0, |
| "tokens/trainable": 116496 |
| }, |
| { |
| "epoch": 1.9, |
| "grad_norm": 0.7480407953262329, |
| "learning_rate": 6.239234449760766e-06, |
| "loss": 0.10476234555244446, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11045, |
| "step": 399, |
| "tokens/total": 1634304, |
| "tokens/train_per_sec_per_gpu": 19.25, |
| "tokens/trainable": 116920 |
| }, |
| { |
| "epoch": 1.9047619047619047, |
| "grad_norm": 0.43783968687057495, |
| "learning_rate": 6.229665071770336e-06, |
| "loss": 0.0689469501376152, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07138, |
| "step": 400, |
| "tokens/total": 1638400, |
| "tokens/train_per_sec_per_gpu": 15.71, |
| "tokens/trainable": 117238 |
| }, |
| { |
| "epoch": 1.9095238095238094, |
| "grad_norm": 1.0542603731155396, |
| "learning_rate": 6.220095693779905e-06, |
| "loss": 0.07620616257190704, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07919, |
| "step": 401, |
| "tokens/total": 1642496, |
| "tokens/train_per_sec_per_gpu": 13.37, |
| "tokens/trainable": 117524 |
| }, |
| { |
| "epoch": 1.9142857142857141, |
| "grad_norm": 0.6692041158676147, |
| "learning_rate": 6.2105263157894745e-06, |
| "loss": 0.07377033680677414, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07656, |
| "step": 402, |
| "tokens/total": 1646592, |
| "tokens/train_per_sec_per_gpu": 15.54, |
| "tokens/trainable": 117841 |
| }, |
| { |
| "epoch": 1.919047619047619, |
| "grad_norm": 1.2519358396530151, |
| "learning_rate": 6.200956937799043e-06, |
| "loss": 0.07229935377836227, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07498, |
| "step": 403, |
| "tokens/total": 1650688, |
| "tokens/train_per_sec_per_gpu": 12.53, |
| "tokens/trainable": 118146 |
| }, |
| { |
| "epoch": 1.9238095238095239, |
| "grad_norm": 0.8904070854187012, |
| "learning_rate": 6.191387559808612e-06, |
| "loss": 0.12588144838809967, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13415, |
| "step": 404, |
| "tokens/total": 1654784, |
| "tokens/train_per_sec_per_gpu": 15.55, |
| "tokens/trainable": 118459 |
| }, |
| { |
| "epoch": 1.9285714285714286, |
| "grad_norm": 0.43530920147895813, |
| "learning_rate": 6.181818181818182e-06, |
| "loss": 0.048857010900974274, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05007, |
| "step": 405, |
| "tokens/total": 1658880, |
| "tokens/train_per_sec_per_gpu": 8.11, |
| "tokens/trainable": 118719 |
| }, |
| { |
| "epoch": 1.9333333333333333, |
| "grad_norm": 1.0130807161331177, |
| "learning_rate": 6.172248803827752e-06, |
| "loss": 0.12293189764022827, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13081, |
| "step": 406, |
| "tokens/total": 1662976, |
| "tokens/train_per_sec_per_gpu": 16.9, |
| "tokens/trainable": 119130 |
| }, |
| { |
| "epoch": 1.938095238095238, |
| "grad_norm": 0.8286269307136536, |
| "learning_rate": 6.162679425837321e-06, |
| "loss": 0.13303330540657043, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14229, |
| "step": 407, |
| "tokens/total": 1667072, |
| "tokens/train_per_sec_per_gpu": 12.2, |
| "tokens/trainable": 119425 |
| }, |
| { |
| "epoch": 1.9428571428571428, |
| "grad_norm": 0.6442462801933289, |
| "learning_rate": 6.153110047846891e-06, |
| "loss": 0.10066962987184525, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10591, |
| "step": 408, |
| "tokens/total": 1671168, |
| "tokens/train_per_sec_per_gpu": 13.14, |
| "tokens/trainable": 119695 |
| }, |
| { |
| "epoch": 1.9476190476190476, |
| "grad_norm": 0.6401284337043762, |
| "learning_rate": 6.14354066985646e-06, |
| "loss": 0.08170831948518753, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.08514, |
| "step": 409, |
| "tokens/total": 1675264, |
| "tokens/train_per_sec_per_gpu": 15.75, |
| "tokens/trainable": 120035 |
| }, |
| { |
| "epoch": 1.9523809523809523, |
| "grad_norm": 0.8274050951004028, |
| "learning_rate": 6.133971291866029e-06, |
| "loss": 0.11197835206985474, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11849, |
| "step": 410, |
| "tokens/total": 1679360, |
| "tokens/train_per_sec_per_gpu": 10.75, |
| "tokens/trainable": 120252 |
| }, |
| { |
| "epoch": 1.9571428571428573, |
| "grad_norm": 0.43950921297073364, |
| "learning_rate": 6.124401913875598e-06, |
| "loss": 0.04837172105908394, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04956, |
| "step": 411, |
| "tokens/total": 1683456, |
| "tokens/train_per_sec_per_gpu": 14.29, |
| "tokens/trainable": 120547 |
| }, |
| { |
| "epoch": 1.961904761904762, |
| "grad_norm": 0.612568199634552, |
| "learning_rate": 6.1148325358851675e-06, |
| "loss": 0.08096243441104889, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08433, |
| "step": 412, |
| "tokens/total": 1687552, |
| "tokens/train_per_sec_per_gpu": 10.18, |
| "tokens/trainable": 120799 |
| }, |
| { |
| "epoch": 1.9666666666666668, |
| "grad_norm": 0.8107662796974182, |
| "learning_rate": 6.105263157894738e-06, |
| "loss": 0.09596385806798935, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10072, |
| "step": 413, |
| "tokens/total": 1691648, |
| "tokens/train_per_sec_per_gpu": 12.37, |
| "tokens/trainable": 121056 |
| }, |
| { |
| "epoch": 1.9714285714285715, |
| "grad_norm": 1.2137329578399658, |
| "learning_rate": 6.095693779904307e-06, |
| "loss": 0.2036145031452179, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.22583, |
| "step": 414, |
| "tokens/total": 1695744, |
| "tokens/train_per_sec_per_gpu": 12.99, |
| "tokens/trainable": 121343 |
| }, |
| { |
| "epoch": 1.9761904761904763, |
| "grad_norm": 0.6540486812591553, |
| "learning_rate": 6.0861244019138765e-06, |
| "loss": 0.11676246672868729, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.12385, |
| "step": 415, |
| "tokens/total": 1699840, |
| "tokens/train_per_sec_per_gpu": 18.55, |
| "tokens/trainable": 121714 |
| }, |
| { |
| "epoch": 1.980952380952381, |
| "grad_norm": 1.0280195474624634, |
| "learning_rate": 6.076555023923445e-06, |
| "loss": 0.15656080842018127, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16948, |
| "step": 416, |
| "tokens/total": 1703936, |
| "tokens/train_per_sec_per_gpu": 12.23, |
| "tokens/trainable": 121998 |
| }, |
| { |
| "epoch": 1.9857142857142858, |
| "grad_norm": 0.7349785566329956, |
| "learning_rate": 6.0669856459330144e-06, |
| "loss": 0.055647723376750946, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05723, |
| "step": 417, |
| "tokens/total": 1708032, |
| "tokens/train_per_sec_per_gpu": 12.44, |
| "tokens/trainable": 122242 |
| }, |
| { |
| "epoch": 1.9904761904761905, |
| "grad_norm": 0.7874380350112915, |
| "learning_rate": 6.057416267942584e-06, |
| "loss": 0.11118850111961365, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11761, |
| "step": 418, |
| "tokens/total": 1712128, |
| "tokens/train_per_sec_per_gpu": 14.29, |
| "tokens/trainable": 122528 |
| }, |
| { |
| "epoch": 1.9952380952380953, |
| "grad_norm": 0.8332106471061707, |
| "learning_rate": 6.047846889952153e-06, |
| "loss": 0.08272654563188553, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08624, |
| "step": 419, |
| "tokens/total": 1716224, |
| "tokens/train_per_sec_per_gpu": 10.18, |
| "tokens/trainable": 122788 |
| }, |
| { |
| "epoch": 2.0, |
| "grad_norm": 0.5468876957893372, |
| "learning_rate": 6.0382775119617234e-06, |
| "loss": 0.0967540517449379, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10159, |
| "step": 420, |
| "tokens/total": 1720320, |
| "tokens/train_per_sec_per_gpu": 13.94, |
| "tokens/trainable": 123212 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_loss": 0.09063497930765152, |
| "eval_ppl": 1.09487, |
| "eval_runtime": 10.3365, |
| "eval_samples_per_second": 8.61, |
| "eval_steps_per_second": 2.225, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 35.58, |
| "memory/max_allocated (GiB)": 35.58, |
| "step": 420 |
| }, |
| { |
| "epoch": 2.0047619047619047, |
| "grad_norm": 0.7299656867980957, |
| "learning_rate": 6.028708133971293e-06, |
| "loss": 0.08038073778152466, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0837, |
| "step": 421, |
| "tokens/total": 1724416, |
| "tokens/train_per_sec_per_gpu": 14.47, |
| "tokens/trainable": 123549 |
| }, |
| { |
| "epoch": 2.0095238095238095, |
| "grad_norm": 0.5367144346237183, |
| "learning_rate": 6.019138755980862e-06, |
| "loss": 0.04271767660975456, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04364, |
| "step": 422, |
| "tokens/total": 1728512, |
| "tokens/train_per_sec_per_gpu": 12.83, |
| "tokens/trainable": 123844 |
| }, |
| { |
| "epoch": 2.0142857142857142, |
| "grad_norm": 0.7927054166793823, |
| "learning_rate": 6.009569377990431e-06, |
| "loss": 0.11457313597202301, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12139, |
| "step": 423, |
| "tokens/total": 1732608, |
| "tokens/train_per_sec_per_gpu": 11.61, |
| "tokens/trainable": 124091 |
| }, |
| { |
| "epoch": 2.019047619047619, |
| "grad_norm": 0.570366382598877, |
| "learning_rate": 6e-06, |
| "loss": 0.06566718965768814, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06787, |
| "step": 424, |
| "tokens/total": 1736704, |
| "tokens/train_per_sec_per_gpu": 19.63, |
| "tokens/trainable": 124430 |
| }, |
| { |
| "epoch": 2.0238095238095237, |
| "grad_norm": 0.6495265960693359, |
| "learning_rate": 5.9904306220095695e-06, |
| "loss": 0.07726883143186569, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08033, |
| "step": 425, |
| "tokens/total": 1740800, |
| "tokens/train_per_sec_per_gpu": 9.77, |
| "tokens/trainable": 124705 |
| }, |
| { |
| "epoch": 2.0285714285714285, |
| "grad_norm": 0.272966593503952, |
| "learning_rate": 5.98086124401914e-06, |
| "loss": 0.033999744802713394, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.03458, |
| "step": 426, |
| "tokens/total": 1744896, |
| "tokens/train_per_sec_per_gpu": 13.54, |
| "tokens/trainable": 125027 |
| }, |
| { |
| "epoch": 2.033333333333333, |
| "grad_norm": 0.5399532318115234, |
| "learning_rate": 5.971291866028709e-06, |
| "loss": 0.0716332346200943, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07426, |
| "step": 427, |
| "tokens/total": 1748992, |
| "tokens/train_per_sec_per_gpu": 16.2, |
| "tokens/trainable": 125408 |
| }, |
| { |
| "epoch": 2.038095238095238, |
| "grad_norm": 0.6029632687568665, |
| "learning_rate": 5.9617224880382786e-06, |
| "loss": 0.09182186424732208, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.09617, |
| "step": 428, |
| "tokens/total": 1753088, |
| "tokens/train_per_sec_per_gpu": 23.19, |
| "tokens/trainable": 125758 |
| }, |
| { |
| "epoch": 2.0428571428571427, |
| "grad_norm": 4.435783386230469, |
| "learning_rate": 5.952153110047847e-06, |
| "loss": 0.11020708829164505, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11651, |
| "step": 429, |
| "tokens/total": 1757184, |
| "tokens/train_per_sec_per_gpu": 13.56, |
| "tokens/trainable": 126085 |
| }, |
| { |
| "epoch": 2.0476190476190474, |
| "grad_norm": 0.5012717247009277, |
| "learning_rate": 5.9425837320574165e-06, |
| "loss": 0.08536789566278458, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08912, |
| "step": 430, |
| "tokens/total": 1761280, |
| "tokens/train_per_sec_per_gpu": 14.74, |
| "tokens/trainable": 126337 |
| }, |
| { |
| "epoch": 2.052380952380952, |
| "grad_norm": 0.6422252058982849, |
| "learning_rate": 5.933014354066986e-06, |
| "loss": 0.08672700822353363, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0906, |
| "step": 431, |
| "tokens/total": 1765376, |
| "tokens/train_per_sec_per_gpu": 14.84, |
| "tokens/trainable": 126681 |
| }, |
| { |
| "epoch": 2.057142857142857, |
| "grad_norm": 0.706159770488739, |
| "learning_rate": 5.923444976076555e-06, |
| "loss": 0.07219837605953217, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07487, |
| "step": 432, |
| "tokens/total": 1769472, |
| "tokens/train_per_sec_per_gpu": 11.37, |
| "tokens/trainable": 126899 |
| }, |
| { |
| "epoch": 2.0619047619047617, |
| "grad_norm": 0.47052693367004395, |
| "learning_rate": 5.9138755980861255e-06, |
| "loss": 0.05155189335346222, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0529, |
| "step": 433, |
| "tokens/total": 1773568, |
| "tokens/train_per_sec_per_gpu": 14.06, |
| "tokens/trainable": 127199 |
| }, |
| { |
| "epoch": 2.066666666666667, |
| "grad_norm": 1.336130142211914, |
| "learning_rate": 5.904306220095695e-06, |
| "loss": 0.09528717398643494, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09997, |
| "step": 434, |
| "tokens/total": 1777664, |
| "tokens/train_per_sec_per_gpu": 15.0, |
| "tokens/trainable": 127465 |
| }, |
| { |
| "epoch": 2.0714285714285716, |
| "grad_norm": 0.7233858108520508, |
| "learning_rate": 5.8947368421052634e-06, |
| "loss": 0.09852813929319382, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10355, |
| "step": 435, |
| "tokens/total": 1781760, |
| "tokens/train_per_sec_per_gpu": 15.27, |
| "tokens/trainable": 127754 |
| }, |
| { |
| "epoch": 2.0761904761904764, |
| "grad_norm": 0.7493842244148254, |
| "learning_rate": 5.885167464114833e-06, |
| "loss": 0.08514859527349472, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08888, |
| "step": 436, |
| "tokens/total": 1785856, |
| "tokens/train_per_sec_per_gpu": 9.27, |
| "tokens/trainable": 128063 |
| }, |
| { |
| "epoch": 2.080952380952381, |
| "grad_norm": 0.4522424340248108, |
| "learning_rate": 5.875598086124402e-06, |
| "loss": 0.05524425208568573, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0568, |
| "step": 437, |
| "tokens/total": 1789952, |
| "tokens/train_per_sec_per_gpu": 12.29, |
| "tokens/trainable": 128372 |
| }, |
| { |
| "epoch": 2.085714285714286, |
| "grad_norm": 0.6085591912269592, |
| "learning_rate": 5.866028708133972e-06, |
| "loss": 0.07435575127601624, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07719, |
| "step": 438, |
| "tokens/total": 1794048, |
| "tokens/train_per_sec_per_gpu": 14.19, |
| "tokens/trainable": 128705 |
| }, |
| { |
| "epoch": 2.0904761904761906, |
| "grad_norm": 0.7386507391929626, |
| "learning_rate": 5.856459330143541e-06, |
| "loss": 0.07269930839538574, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07541, |
| "step": 439, |
| "tokens/total": 1798144, |
| "tokens/train_per_sec_per_gpu": 19.96, |
| "tokens/trainable": 129037 |
| }, |
| { |
| "epoch": 2.0952380952380953, |
| "grad_norm": 0.6475787162780762, |
| "learning_rate": 5.846889952153111e-06, |
| "loss": 0.06826756149530411, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07065, |
| "step": 440, |
| "tokens/total": 1802240, |
| "tokens/train_per_sec_per_gpu": 11.72, |
| "tokens/trainable": 129310 |
| }, |
| { |
| "epoch": 2.1, |
| "grad_norm": 0.5135707855224609, |
| "learning_rate": 5.837320574162681e-06, |
| "loss": 0.0561576709151268, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.05776, |
| "step": 441, |
| "tokens/total": 1806336, |
| "tokens/train_per_sec_per_gpu": 12.27, |
| "tokens/trainable": 129587 |
| }, |
| { |
| "epoch": 2.104761904761905, |
| "grad_norm": 0.5076371431350708, |
| "learning_rate": 5.827751196172249e-06, |
| "loss": 0.09091908484697342, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09518, |
| "step": 442, |
| "tokens/total": 1810432, |
| "tokens/train_per_sec_per_gpu": 13.89, |
| "tokens/trainable": 129851 |
| }, |
| { |
| "epoch": 2.1095238095238096, |
| "grad_norm": 0.7993151545524597, |
| "learning_rate": 5.8181818181818185e-06, |
| "loss": 0.11744911223649979, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12462, |
| "step": 443, |
| "tokens/total": 1814528, |
| "tokens/train_per_sec_per_gpu": 14.06, |
| "tokens/trainable": 130137 |
| }, |
| { |
| "epoch": 2.1142857142857143, |
| "grad_norm": 0.6741457581520081, |
| "learning_rate": 5.808612440191388e-06, |
| "loss": 0.06119607016444206, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06311, |
| "step": 444, |
| "tokens/total": 1818624, |
| "tokens/train_per_sec_per_gpu": 9.3, |
| "tokens/trainable": 130370 |
| }, |
| { |
| "epoch": 2.119047619047619, |
| "grad_norm": 0.4995241165161133, |
| "learning_rate": 5.799043062200957e-06, |
| "loss": 0.060865581035614014, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06276, |
| "step": 445, |
| "tokens/total": 1822720, |
| "tokens/train_per_sec_per_gpu": 13.13, |
| "tokens/trainable": 130620 |
| }, |
| { |
| "epoch": 2.123809523809524, |
| "grad_norm": 1.2829558849334717, |
| "learning_rate": 5.789473684210527e-06, |
| "loss": 0.10799669474363327, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11404, |
| "step": 446, |
| "tokens/total": 1826816, |
| "tokens/train_per_sec_per_gpu": 10.65, |
| "tokens/trainable": 130832 |
| }, |
| { |
| "epoch": 2.1285714285714286, |
| "grad_norm": 0.6178713440895081, |
| "learning_rate": 5.779904306220097e-06, |
| "loss": 0.07918387651443481, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0824, |
| "step": 447, |
| "tokens/total": 1830912, |
| "tokens/train_per_sec_per_gpu": 10.76, |
| "tokens/trainable": 131075 |
| }, |
| { |
| "epoch": 2.1333333333333333, |
| "grad_norm": 0.8874167203903198, |
| "learning_rate": 5.770334928229665e-06, |
| "loss": 0.10994468629360199, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11622, |
| "step": 448, |
| "tokens/total": 1835008, |
| "tokens/train_per_sec_per_gpu": 14.39, |
| "tokens/trainable": 131390 |
| }, |
| { |
| "epoch": 2.138095238095238, |
| "grad_norm": 0.5274553894996643, |
| "learning_rate": 5.760765550239235e-06, |
| "loss": 0.04991922155022621, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05119, |
| "step": 449, |
| "tokens/total": 1839104, |
| "tokens/train_per_sec_per_gpu": 9.87, |
| "tokens/trainable": 131641 |
| }, |
| { |
| "epoch": 2.142857142857143, |
| "grad_norm": 0.5265685319900513, |
| "learning_rate": 5.751196172248804e-06, |
| "loss": 0.06800280511379242, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07037, |
| "step": 450, |
| "tokens/total": 1843200, |
| "tokens/train_per_sec_per_gpu": 18.67, |
| "tokens/trainable": 131988 |
| }, |
| { |
| "epoch": 2.1476190476190475, |
| "grad_norm": 0.6688624620437622, |
| "learning_rate": 5.741626794258374e-06, |
| "loss": 0.07494084537029266, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07782, |
| "step": 451, |
| "tokens/total": 1847296, |
| "tokens/train_per_sec_per_gpu": 10.58, |
| "tokens/trainable": 132268 |
| }, |
| { |
| "epoch": 2.1523809523809523, |
| "grad_norm": 0.7583678960800171, |
| "learning_rate": 5.732057416267943e-06, |
| "loss": 0.07973390817642212, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.083, |
| "step": 452, |
| "tokens/total": 1851392, |
| "tokens/train_per_sec_per_gpu": 12.94, |
| "tokens/trainable": 132537 |
| }, |
| { |
| "epoch": 2.157142857142857, |
| "grad_norm": 0.7710968255996704, |
| "learning_rate": 5.722488038277512e-06, |
| "loss": 0.059347622096538544, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06114, |
| "step": 453, |
| "tokens/total": 1855488, |
| "tokens/train_per_sec_per_gpu": 8.91, |
| "tokens/trainable": 132725 |
| }, |
| { |
| "epoch": 2.1619047619047618, |
| "grad_norm": 0.6877778768539429, |
| "learning_rate": 5.712918660287083e-06, |
| "loss": 0.0703950971364975, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07293, |
| "step": 454, |
| "tokens/total": 1859584, |
| "tokens/train_per_sec_per_gpu": 13.66, |
| "tokens/trainable": 132966 |
| }, |
| { |
| "epoch": 2.1666666666666665, |
| "grad_norm": 0.9029383063316345, |
| "learning_rate": 5.70334928229665e-06, |
| "loss": 0.1015714481472969, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10691, |
| "step": 455, |
| "tokens/total": 1863680, |
| "tokens/train_per_sec_per_gpu": 9.51, |
| "tokens/trainable": 133230 |
| }, |
| { |
| "epoch": 2.1714285714285713, |
| "grad_norm": 0.68222975730896, |
| "learning_rate": 5.693779904306221e-06, |
| "loss": 0.08985096216201782, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09401, |
| "step": 456, |
| "tokens/total": 1867776, |
| "tokens/train_per_sec_per_gpu": 23.65, |
| "tokens/trainable": 133595 |
| }, |
| { |
| "epoch": 2.176190476190476, |
| "grad_norm": 0.8007204532623291, |
| "learning_rate": 5.68421052631579e-06, |
| "loss": 0.0790681391954422, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08228, |
| "step": 457, |
| "tokens/total": 1871872, |
| "tokens/train_per_sec_per_gpu": 18.34, |
| "tokens/trainable": 133904 |
| }, |
| { |
| "epoch": 2.1809523809523808, |
| "grad_norm": 0.46541061997413635, |
| "learning_rate": 5.674641148325359e-06, |
| "loss": 0.049216851592063904, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05045, |
| "step": 458, |
| "tokens/total": 1875968, |
| "tokens/train_per_sec_per_gpu": 8.31, |
| "tokens/trainable": 134132 |
| }, |
| { |
| "epoch": 2.185714285714286, |
| "grad_norm": 0.4131176173686981, |
| "learning_rate": 5.665071770334929e-06, |
| "loss": 0.0468255877494812, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04794, |
| "step": 459, |
| "tokens/total": 1880064, |
| "tokens/train_per_sec_per_gpu": 17.36, |
| "tokens/trainable": 134429 |
| }, |
| { |
| "epoch": 2.1904761904761907, |
| "grad_norm": 0.7962276935577393, |
| "learning_rate": 5.655502392344498e-06, |
| "loss": 0.09427708387374878, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09886, |
| "step": 460, |
| "tokens/total": 1884160, |
| "tokens/train_per_sec_per_gpu": 15.21, |
| "tokens/trainable": 134737 |
| }, |
| { |
| "epoch": 2.1952380952380954, |
| "grad_norm": 1.0497944355010986, |
| "learning_rate": 5.645933014354067e-06, |
| "loss": 0.12763239443302155, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13614, |
| "step": 461, |
| "tokens/total": 1888256, |
| "tokens/train_per_sec_per_gpu": 14.99, |
| "tokens/trainable": 135039 |
| }, |
| { |
| "epoch": 2.2, |
| "grad_norm": 0.676685094833374, |
| "learning_rate": 5.636363636363636e-06, |
| "loss": 0.1088184341788292, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11496, |
| "step": 462, |
| "tokens/total": 1892352, |
| "tokens/train_per_sec_per_gpu": 13.36, |
| "tokens/trainable": 135338 |
| }, |
| { |
| "epoch": 2.204761904761905, |
| "grad_norm": 0.8243446946144104, |
| "learning_rate": 5.626794258373206e-06, |
| "loss": 0.08344703912734985, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08703, |
| "step": 463, |
| "tokens/total": 1896448, |
| "tokens/train_per_sec_per_gpu": 16.53, |
| "tokens/trainable": 135645 |
| }, |
| { |
| "epoch": 2.2095238095238097, |
| "grad_norm": 0.48808151483535767, |
| "learning_rate": 5.617224880382776e-06, |
| "loss": 0.04418756067752838, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04518, |
| "step": 464, |
| "tokens/total": 1900544, |
| "tokens/train_per_sec_per_gpu": 12.72, |
| "tokens/trainable": 135877 |
| }, |
| { |
| "epoch": 2.2142857142857144, |
| "grad_norm": 0.8586388230323792, |
| "learning_rate": 5.607655502392345e-06, |
| "loss": 0.0482182651758194, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0494, |
| "step": 465, |
| "tokens/total": 1904640, |
| "tokens/train_per_sec_per_gpu": 12.92, |
| "tokens/trainable": 136106 |
| }, |
| { |
| "epoch": 2.219047619047619, |
| "grad_norm": 1.033241868019104, |
| "learning_rate": 5.5980861244019145e-06, |
| "loss": 0.13134968280792236, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14037, |
| "step": 466, |
| "tokens/total": 1908736, |
| "tokens/train_per_sec_per_gpu": 18.94, |
| "tokens/trainable": 136508 |
| }, |
| { |
| "epoch": 2.223809523809524, |
| "grad_norm": 0.6962218880653381, |
| "learning_rate": 5.588516746411484e-06, |
| "loss": 0.1050298735499382, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11074, |
| "step": 467, |
| "tokens/total": 1912832, |
| "tokens/train_per_sec_per_gpu": 19.56, |
| "tokens/trainable": 136895 |
| }, |
| { |
| "epoch": 2.2285714285714286, |
| "grad_norm": 1.0098704099655151, |
| "learning_rate": 5.578947368421052e-06, |
| "loss": 0.10072275996208191, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10597, |
| "step": 468, |
| "tokens/total": 1916928, |
| "tokens/train_per_sec_per_gpu": 7.89, |
| "tokens/trainable": 137091 |
| }, |
| { |
| "epoch": 2.2333333333333334, |
| "grad_norm": 0.6704676151275635, |
| "learning_rate": 5.569377990430623e-06, |
| "loss": 0.059372853487730026, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06117, |
| "step": 469, |
| "tokens/total": 1921024, |
| "tokens/train_per_sec_per_gpu": 16.83, |
| "tokens/trainable": 137444 |
| }, |
| { |
| "epoch": 2.238095238095238, |
| "grad_norm": 0.6171624064445496, |
| "learning_rate": 5.559808612440192e-06, |
| "loss": 0.06646374613046646, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06872, |
| "step": 470, |
| "tokens/total": 1925120, |
| "tokens/train_per_sec_per_gpu": 13.77, |
| "tokens/trainable": 137741 |
| }, |
| { |
| "epoch": 2.242857142857143, |
| "grad_norm": 1.2899796962738037, |
| "learning_rate": 5.550239234449761e-06, |
| "loss": 0.1084589809179306, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11456, |
| "step": 471, |
| "tokens/total": 1929216, |
| "tokens/train_per_sec_per_gpu": 10.58, |
| "tokens/trainable": 137958 |
| }, |
| { |
| "epoch": 2.2476190476190476, |
| "grad_norm": 0.7799538969993591, |
| "learning_rate": 5.540669856459331e-06, |
| "loss": 0.08141331374645233, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08482, |
| "step": 472, |
| "tokens/total": 1933312, |
| "tokens/train_per_sec_per_gpu": 10.94, |
| "tokens/trainable": 138201 |
| }, |
| { |
| "epoch": 2.2523809523809524, |
| "grad_norm": 1.5501976013183594, |
| "learning_rate": 5.5311004784689e-06, |
| "loss": 0.05066931992769241, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05197, |
| "step": 473, |
| "tokens/total": 1937408, |
| "tokens/train_per_sec_per_gpu": 10.48, |
| "tokens/trainable": 138531 |
| }, |
| { |
| "epoch": 2.257142857142857, |
| "grad_norm": 0.6510878205299377, |
| "learning_rate": 5.521531100478469e-06, |
| "loss": 0.0661131888628006, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.06835, |
| "step": 474, |
| "tokens/total": 1941504, |
| "tokens/train_per_sec_per_gpu": 15.26, |
| "tokens/trainable": 138862 |
| }, |
| { |
| "epoch": 2.261904761904762, |
| "grad_norm": 0.5600479245185852, |
| "learning_rate": 5.511961722488038e-06, |
| "loss": 0.07156892865896225, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07419, |
| "step": 475, |
| "tokens/total": 1945600, |
| "tokens/train_per_sec_per_gpu": 14.49, |
| "tokens/trainable": 139227 |
| }, |
| { |
| "epoch": 2.2666666666666666, |
| "grad_norm": 0.22503618896007538, |
| "learning_rate": 5.502392344497608e-06, |
| "loss": 0.024924561381340027, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02524, |
| "step": 476, |
| "tokens/total": 1949696, |
| "tokens/train_per_sec_per_gpu": 17.97, |
| "tokens/trainable": 139524 |
| }, |
| { |
| "epoch": 2.2714285714285714, |
| "grad_norm": 11.384724617004395, |
| "learning_rate": 5.492822966507178e-06, |
| "loss": 0.08557814359664917, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08935, |
| "step": 477, |
| "tokens/total": 1953792, |
| "tokens/train_per_sec_per_gpu": 14.9, |
| "tokens/trainable": 139801 |
| }, |
| { |
| "epoch": 2.276190476190476, |
| "grad_norm": 0.8630121350288391, |
| "learning_rate": 5.483253588516747e-06, |
| "loss": 0.07430025935173035, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07713, |
| "step": 478, |
| "tokens/total": 1957888, |
| "tokens/train_per_sec_per_gpu": 13.43, |
| "tokens/trainable": 140096 |
| }, |
| { |
| "epoch": 2.280952380952381, |
| "grad_norm": 0.5071492195129395, |
| "learning_rate": 5.4736842105263165e-06, |
| "loss": 0.06158454343676567, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06352, |
| "step": 479, |
| "tokens/total": 1961984, |
| "tokens/train_per_sec_per_gpu": 13.16, |
| "tokens/trainable": 140333 |
| }, |
| { |
| "epoch": 2.2857142857142856, |
| "grad_norm": 0.95302414894104, |
| "learning_rate": 5.464114832535886e-06, |
| "loss": 0.10877791047096252, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11491, |
| "step": 480, |
| "tokens/total": 1966080, |
| "tokens/train_per_sec_per_gpu": 20.9, |
| "tokens/trainable": 140702 |
| }, |
| { |
| "epoch": 2.2904761904761903, |
| "grad_norm": 0.7657179236412048, |
| "learning_rate": 5.4545454545454545e-06, |
| "loss": 0.08901463449001312, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0931, |
| "step": 481, |
| "tokens/total": 1970176, |
| "tokens/train_per_sec_per_gpu": 16.74, |
| "tokens/trainable": 141092 |
| }, |
| { |
| "epoch": 2.295238095238095, |
| "grad_norm": 0.7546993494033813, |
| "learning_rate": 5.444976076555024e-06, |
| "loss": 0.11137181520462036, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11781, |
| "step": 482, |
| "tokens/total": 1974272, |
| "tokens/train_per_sec_per_gpu": 11.91, |
| "tokens/trainable": 141376 |
| }, |
| { |
| "epoch": 2.3, |
| "grad_norm": 0.9208468198776245, |
| "learning_rate": 5.435406698564594e-06, |
| "loss": 0.08001670241355896, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08331, |
| "step": 483, |
| "tokens/total": 1978368, |
| "tokens/train_per_sec_per_gpu": 11.28, |
| "tokens/trainable": 141667 |
| }, |
| { |
| "epoch": 2.3047619047619046, |
| "grad_norm": 0.7315728664398193, |
| "learning_rate": 5.4258373205741635e-06, |
| "loss": 0.09134405851364136, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09565, |
| "step": 484, |
| "tokens/total": 1982464, |
| "tokens/train_per_sec_per_gpu": 12.09, |
| "tokens/trainable": 142026 |
| }, |
| { |
| "epoch": 2.3095238095238093, |
| "grad_norm": 1.0361758470535278, |
| "learning_rate": 5.416267942583733e-06, |
| "loss": 0.09037132561206818, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09458, |
| "step": 485, |
| "tokens/total": 1986560, |
| "tokens/train_per_sec_per_gpu": 10.34, |
| "tokens/trainable": 142257 |
| }, |
| { |
| "epoch": 2.314285714285714, |
| "grad_norm": 0.9890063405036926, |
| "learning_rate": 5.406698564593302e-06, |
| "loss": 0.0984952449798584, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10351, |
| "step": 486, |
| "tokens/total": 1990656, |
| "tokens/train_per_sec_per_gpu": 14.67, |
| "tokens/trainable": 142556 |
| }, |
| { |
| "epoch": 2.319047619047619, |
| "grad_norm": 0.9803997874259949, |
| "learning_rate": 5.397129186602871e-06, |
| "loss": 0.08330434560775757, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08687, |
| "step": 487, |
| "tokens/total": 1994752, |
| "tokens/train_per_sec_per_gpu": 13.84, |
| "tokens/trainable": 142824 |
| }, |
| { |
| "epoch": 2.323809523809524, |
| "grad_norm": 0.8348151445388794, |
| "learning_rate": 5.38755980861244e-06, |
| "loss": 0.09529228508472443, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09998, |
| "step": 488, |
| "tokens/total": 1998848, |
| "tokens/train_per_sec_per_gpu": 15.12, |
| "tokens/trainable": 143109 |
| }, |
| { |
| "epoch": 2.3285714285714287, |
| "grad_norm": 1.044250249862671, |
| "learning_rate": 5.3779904306220096e-06, |
| "loss": 0.10967185348272324, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11591, |
| "step": 489, |
| "tokens/total": 2002944, |
| "tokens/train_per_sec_per_gpu": 11.37, |
| "tokens/trainable": 143371 |
| }, |
| { |
| "epoch": 2.3333333333333335, |
| "grad_norm": 0.7318430542945862, |
| "learning_rate": 5.36842105263158e-06, |
| "loss": 0.061104319989681244, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06301, |
| "step": 490, |
| "tokens/total": 2007040, |
| "tokens/train_per_sec_per_gpu": 12.7, |
| "tokens/trainable": 143648 |
| }, |
| { |
| "epoch": 2.3380952380952382, |
| "grad_norm": 0.6046728491783142, |
| "learning_rate": 5.358851674641149e-06, |
| "loss": 0.07590795308351517, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07886, |
| "step": 491, |
| "tokens/total": 2011136, |
| "tokens/train_per_sec_per_gpu": 11.62, |
| "tokens/trainable": 143964 |
| }, |
| { |
| "epoch": 2.342857142857143, |
| "grad_norm": 0.36420607566833496, |
| "learning_rate": 5.3492822966507186e-06, |
| "loss": 0.04336003586649895, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04431, |
| "step": 492, |
| "tokens/total": 2015232, |
| "tokens/train_per_sec_per_gpu": 12.39, |
| "tokens/trainable": 144271 |
| }, |
| { |
| "epoch": 2.3476190476190477, |
| "grad_norm": 0.6114564538002014, |
| "learning_rate": 5.339712918660288e-06, |
| "loss": 0.06792014837265015, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07028, |
| "step": 493, |
| "tokens/total": 2019328, |
| "tokens/train_per_sec_per_gpu": 15.04, |
| "tokens/trainable": 144580 |
| }, |
| { |
| "epoch": 2.3523809523809525, |
| "grad_norm": 0.49948400259017944, |
| "learning_rate": 5.3301435406698565e-06, |
| "loss": 0.050412971526384354, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05171, |
| "step": 494, |
| "tokens/total": 2023424, |
| "tokens/train_per_sec_per_gpu": 11.16, |
| "tokens/trainable": 144853 |
| }, |
| { |
| "epoch": 2.357142857142857, |
| "grad_norm": 1.2982183694839478, |
| "learning_rate": 5.320574162679426e-06, |
| "loss": 0.09459419548511505, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09921, |
| "step": 495, |
| "tokens/total": 2027520, |
| "tokens/train_per_sec_per_gpu": 8.11, |
| "tokens/trainable": 145057 |
| }, |
| { |
| "epoch": 2.361904761904762, |
| "grad_norm": 1.496275544166565, |
| "learning_rate": 5.311004784688995e-06, |
| "loss": 0.13228096067905426, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14143, |
| "step": 496, |
| "tokens/total": 2031616, |
| "tokens/train_per_sec_per_gpu": 11.86, |
| "tokens/trainable": 145299 |
| }, |
| { |
| "epoch": 2.3666666666666667, |
| "grad_norm": 0.8566056489944458, |
| "learning_rate": 5.3014354066985655e-06, |
| "loss": 0.09681127220392227, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10165, |
| "step": 497, |
| "tokens/total": 2035712, |
| "tokens/train_per_sec_per_gpu": 11.2, |
| "tokens/trainable": 145532 |
| }, |
| { |
| "epoch": 2.3714285714285714, |
| "grad_norm": 0.8557098507881165, |
| "learning_rate": 5.291866028708135e-06, |
| "loss": 0.14163172245025635, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15215, |
| "step": 498, |
| "tokens/total": 2039808, |
| "tokens/train_per_sec_per_gpu": 26.72, |
| "tokens/trainable": 145990 |
| }, |
| { |
| "epoch": 2.376190476190476, |
| "grad_norm": 0.6874766945838928, |
| "learning_rate": 5.282296650717704e-06, |
| "loss": 0.0928252637386322, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09727, |
| "step": 499, |
| "tokens/total": 2043904, |
| "tokens/train_per_sec_per_gpu": 16.58, |
| "tokens/trainable": 146338 |
| }, |
| { |
| "epoch": 2.380952380952381, |
| "grad_norm": 0.5962072014808655, |
| "learning_rate": 5.272727272727273e-06, |
| "loss": 0.045713819563388824, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04677, |
| "step": 500, |
| "tokens/total": 2048000, |
| "tokens/train_per_sec_per_gpu": 9.35, |
| "tokens/trainable": 146621 |
| }, |
| { |
| "epoch": 2.3857142857142857, |
| "grad_norm": 0.6994227766990662, |
| "learning_rate": 5.263157894736842e-06, |
| "loss": 0.0747811570763588, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07765, |
| "step": 501, |
| "tokens/total": 2052096, |
| "tokens/train_per_sec_per_gpu": 10.56, |
| "tokens/trainable": 146854 |
| }, |
| { |
| "epoch": 2.3904761904761904, |
| "grad_norm": 0.7604936361312866, |
| "learning_rate": 5.253588516746412e-06, |
| "loss": 0.07655870169401169, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07957, |
| "step": 502, |
| "tokens/total": 2056192, |
| "tokens/train_per_sec_per_gpu": 11.09, |
| "tokens/trainable": 147075 |
| }, |
| { |
| "epoch": 2.395238095238095, |
| "grad_norm": 0.9588606357574463, |
| "learning_rate": 5.244019138755981e-06, |
| "loss": 0.11817756295204163, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12544, |
| "step": 503, |
| "tokens/total": 2060288, |
| "tokens/train_per_sec_per_gpu": 12.14, |
| "tokens/trainable": 147362 |
| }, |
| { |
| "epoch": 2.4, |
| "grad_norm": 0.6628692746162415, |
| "learning_rate": 5.234449760765551e-06, |
| "loss": 0.09815141558647156, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10313, |
| "step": 504, |
| "tokens/total": 2064384, |
| "tokens/train_per_sec_per_gpu": 22.65, |
| "tokens/trainable": 147733 |
| }, |
| { |
| "epoch": 2.4047619047619047, |
| "grad_norm": 1.0721534490585327, |
| "learning_rate": 5.224880382775121e-06, |
| "loss": 0.1393708884716034, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.14955, |
| "step": 505, |
| "tokens/total": 2068480, |
| "tokens/train_per_sec_per_gpu": 12.26, |
| "tokens/trainable": 148046 |
| }, |
| { |
| "epoch": 2.4095238095238094, |
| "grad_norm": 0.789513349533081, |
| "learning_rate": 5.215311004784689e-06, |
| "loss": 0.13143616914749146, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14047, |
| "step": 506, |
| "tokens/total": 2072576, |
| "tokens/train_per_sec_per_gpu": 15.5, |
| "tokens/trainable": 148318 |
| }, |
| { |
| "epoch": 2.414285714285714, |
| "grad_norm": 0.5102568864822388, |
| "learning_rate": 5.2057416267942586e-06, |
| "loss": 0.044160425662994385, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04515, |
| "step": 507, |
| "tokens/total": 2076672, |
| "tokens/train_per_sec_per_gpu": 10.74, |
| "tokens/trainable": 148594 |
| }, |
| { |
| "epoch": 2.419047619047619, |
| "grad_norm": 0.717756986618042, |
| "learning_rate": 5.196172248803828e-06, |
| "loss": 0.0987585261464119, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1038, |
| "step": 508, |
| "tokens/total": 2080768, |
| "tokens/train_per_sec_per_gpu": 12.98, |
| "tokens/trainable": 148857 |
| }, |
| { |
| "epoch": 2.4238095238095236, |
| "grad_norm": 0.8132109642028809, |
| "learning_rate": 5.186602870813397e-06, |
| "loss": 0.0786421149969101, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08182, |
| "step": 509, |
| "tokens/total": 2084864, |
| "tokens/train_per_sec_per_gpu": 10.09, |
| "tokens/trainable": 149160 |
| }, |
| { |
| "epoch": 2.4285714285714284, |
| "grad_norm": 1.0510293245315552, |
| "learning_rate": 5.177033492822967e-06, |
| "loss": 0.11246781051158905, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11904, |
| "step": 510, |
| "tokens/total": 2088960, |
| "tokens/train_per_sec_per_gpu": 11.42, |
| "tokens/trainable": 149476 |
| }, |
| { |
| "epoch": 2.4333333333333336, |
| "grad_norm": 0.9610063433647156, |
| "learning_rate": 5.167464114832537e-06, |
| "loss": 0.1295289546251297, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13829, |
| "step": 511, |
| "tokens/total": 2093056, |
| "tokens/train_per_sec_per_gpu": 10.39, |
| "tokens/trainable": 149733 |
| }, |
| { |
| "epoch": 2.4380952380952383, |
| "grad_norm": 0.6407898664474487, |
| "learning_rate": 5.157894736842106e-06, |
| "loss": 0.041559718549251556, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04244, |
| "step": 512, |
| "tokens/total": 2097152, |
| "tokens/train_per_sec_per_gpu": 12.15, |
| "tokens/trainable": 149960 |
| }, |
| { |
| "epoch": 2.442857142857143, |
| "grad_norm": 0.6103262901306152, |
| "learning_rate": 5.148325358851675e-06, |
| "loss": 0.04148872196674347, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04236, |
| "step": 513, |
| "tokens/total": 2101248, |
| "tokens/train_per_sec_per_gpu": 10.17, |
| "tokens/trainable": 150178 |
| }, |
| { |
| "epoch": 2.447619047619048, |
| "grad_norm": 0.4540882110595703, |
| "learning_rate": 5.138755980861244e-06, |
| "loss": 0.0655214786529541, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06772, |
| "step": 514, |
| "tokens/total": 2105344, |
| "tokens/train_per_sec_per_gpu": 16.99, |
| "tokens/trainable": 150512 |
| }, |
| { |
| "epoch": 2.4523809523809526, |
| "grad_norm": 0.7555355429649353, |
| "learning_rate": 5.129186602870814e-06, |
| "loss": 0.11386040598154068, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1206, |
| "step": 515, |
| "tokens/total": 2109440, |
| "tokens/train_per_sec_per_gpu": 19.26, |
| "tokens/trainable": 150978 |
| }, |
| { |
| "epoch": 2.4571428571428573, |
| "grad_norm": 0.7311438322067261, |
| "learning_rate": 5.119617224880383e-06, |
| "loss": 0.0901159942150116, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0943, |
| "step": 516, |
| "tokens/total": 2113536, |
| "tokens/train_per_sec_per_gpu": 10.0, |
| "tokens/trainable": 151213 |
| }, |
| { |
| "epoch": 2.461904761904762, |
| "grad_norm": 0.8779038786888123, |
| "learning_rate": 5.1100478468899524e-06, |
| "loss": 0.08724822103977203, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09117, |
| "step": 517, |
| "tokens/total": 2117632, |
| "tokens/train_per_sec_per_gpu": 11.79, |
| "tokens/trainable": 151536 |
| }, |
| { |
| "epoch": 2.466666666666667, |
| "grad_norm": 0.43343234062194824, |
| "learning_rate": 5.100478468899523e-06, |
| "loss": 0.04299956187605858, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04394, |
| "step": 518, |
| "tokens/total": 2121728, |
| "tokens/train_per_sec_per_gpu": 10.73, |
| "tokens/trainable": 151773 |
| }, |
| { |
| "epoch": 2.4714285714285715, |
| "grad_norm": 0.2579718232154846, |
| "learning_rate": 5.090909090909091e-06, |
| "loss": 0.029250185936689377, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02968, |
| "step": 519, |
| "tokens/total": 2125824, |
| "tokens/train_per_sec_per_gpu": 10.96, |
| "tokens/trainable": 151986 |
| }, |
| { |
| "epoch": 2.4761904761904763, |
| "grad_norm": 0.6281625032424927, |
| "learning_rate": 5.081339712918661e-06, |
| "loss": 0.04683239012956619, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04795, |
| "step": 520, |
| "tokens/total": 2129920, |
| "tokens/train_per_sec_per_gpu": 9.73, |
| "tokens/trainable": 152226 |
| }, |
| { |
| "epoch": 2.480952380952381, |
| "grad_norm": 0.8255984783172607, |
| "learning_rate": 5.07177033492823e-06, |
| "loss": 0.106313556432724, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.11217, |
| "step": 521, |
| "tokens/total": 2134016, |
| "tokens/train_per_sec_per_gpu": 20.19, |
| "tokens/trainable": 152535 |
| }, |
| { |
| "epoch": 2.4857142857142858, |
| "grad_norm": 0.8105759024620056, |
| "learning_rate": 5.062200956937799e-06, |
| "loss": 0.07927460223436356, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0825, |
| "step": 522, |
| "tokens/total": 2138112, |
| "tokens/train_per_sec_per_gpu": 11.63, |
| "tokens/trainable": 152762 |
| }, |
| { |
| "epoch": 2.4904761904761905, |
| "grad_norm": 0.4739595949649811, |
| "learning_rate": 5.052631578947369e-06, |
| "loss": 0.060388147830963135, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06225, |
| "step": 523, |
| "tokens/total": 2142208, |
| "tokens/train_per_sec_per_gpu": 21.16, |
| "tokens/trainable": 153097 |
| }, |
| { |
| "epoch": 2.4952380952380953, |
| "grad_norm": 0.6768450736999512, |
| "learning_rate": 5.043062200956938e-06, |
| "loss": 0.058053936809301376, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.05977, |
| "step": 524, |
| "tokens/total": 2146304, |
| "tokens/train_per_sec_per_gpu": 17.86, |
| "tokens/trainable": 153414 |
| }, |
| { |
| "epoch": 2.5, |
| "grad_norm": 0.7326765656471252, |
| "learning_rate": 5.033492822966508e-06, |
| "loss": 0.10788488388061523, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11392, |
| "step": 525, |
| "tokens/total": 2150400, |
| "tokens/train_per_sec_per_gpu": 11.54, |
| "tokens/trainable": 153759 |
| }, |
| { |
| "epoch": 2.5047619047619047, |
| "grad_norm": 1.1349955797195435, |
| "learning_rate": 5.023923444976077e-06, |
| "loss": 0.15112200379371643, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.16314, |
| "step": 526, |
| "tokens/total": 2154496, |
| "tokens/train_per_sec_per_gpu": 12.13, |
| "tokens/trainable": 154052 |
| }, |
| { |
| "epoch": 2.5095238095238095, |
| "grad_norm": 0.9874095320701599, |
| "learning_rate": 5.014354066985646e-06, |
| "loss": 0.0937184989452362, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09825, |
| "step": 527, |
| "tokens/total": 2158592, |
| "tokens/train_per_sec_per_gpu": 10.86, |
| "tokens/trainable": 154391 |
| }, |
| { |
| "epoch": 2.5142857142857142, |
| "grad_norm": 0.9318037033081055, |
| "learning_rate": 5.004784688995216e-06, |
| "loss": 0.08227807283401489, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08576, |
| "step": 528, |
| "tokens/total": 2162688, |
| "tokens/train_per_sec_per_gpu": 11.81, |
| "tokens/trainable": 154642 |
| }, |
| { |
| "epoch": 2.519047619047619, |
| "grad_norm": 0.5632556080818176, |
| "learning_rate": 4.995215311004785e-06, |
| "loss": 0.05418809875845909, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05568, |
| "step": 529, |
| "tokens/total": 2166784, |
| "tokens/train_per_sec_per_gpu": 13.83, |
| "tokens/trainable": 154953 |
| }, |
| { |
| "epoch": 2.5238095238095237, |
| "grad_norm": 0.8830906748771667, |
| "learning_rate": 4.9856459330143545e-06, |
| "loss": 0.15271666646003723, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.16499, |
| "step": 530, |
| "tokens/total": 2170880, |
| "tokens/train_per_sec_per_gpu": 16.43, |
| "tokens/trainable": 155427 |
| }, |
| { |
| "epoch": 2.5285714285714285, |
| "grad_norm": 0.8743612766265869, |
| "learning_rate": 4.976076555023924e-06, |
| "loss": 0.12847112119197845, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13709, |
| "step": 531, |
| "tokens/total": 2174976, |
| "tokens/train_per_sec_per_gpu": 13.95, |
| "tokens/trainable": 155739 |
| }, |
| { |
| "epoch": 2.533333333333333, |
| "grad_norm": 0.6132271885871887, |
| "learning_rate": 4.966507177033493e-06, |
| "loss": 0.08681163191795349, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09069, |
| "step": 532, |
| "tokens/total": 2179072, |
| "tokens/train_per_sec_per_gpu": 14.1, |
| "tokens/trainable": 156058 |
| }, |
| { |
| "epoch": 2.538095238095238, |
| "grad_norm": 1.1388221979141235, |
| "learning_rate": 4.956937799043063e-06, |
| "loss": 0.11160650104284286, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11807, |
| "step": 533, |
| "tokens/total": 2183168, |
| "tokens/train_per_sec_per_gpu": 10.02, |
| "tokens/trainable": 156291 |
| }, |
| { |
| "epoch": 2.5428571428571427, |
| "grad_norm": 0.5418483018875122, |
| "learning_rate": 4.947368421052632e-06, |
| "loss": 0.06877899914979935, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.0712, |
| "step": 534, |
| "tokens/total": 2187264, |
| "tokens/train_per_sec_per_gpu": 12.37, |
| "tokens/trainable": 156561 |
| }, |
| { |
| "epoch": 2.5476190476190474, |
| "grad_norm": 0.4854298233985901, |
| "learning_rate": 4.9377990430622014e-06, |
| "loss": 0.07671515643596649, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07973, |
| "step": 535, |
| "tokens/total": 2191360, |
| "tokens/train_per_sec_per_gpu": 13.33, |
| "tokens/trainable": 156863 |
| }, |
| { |
| "epoch": 2.552380952380952, |
| "grad_norm": 0.6366608142852783, |
| "learning_rate": 4.928229665071771e-06, |
| "loss": 0.052159231156110764, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05354, |
| "step": 536, |
| "tokens/total": 2195456, |
| "tokens/train_per_sec_per_gpu": 12.15, |
| "tokens/trainable": 157097 |
| }, |
| { |
| "epoch": 2.557142857142857, |
| "grad_norm": 0.8776169419288635, |
| "learning_rate": 4.91866028708134e-06, |
| "loss": 0.09952576458454132, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10465, |
| "step": 537, |
| "tokens/total": 2199552, |
| "tokens/train_per_sec_per_gpu": 14.1, |
| "tokens/trainable": 157357 |
| }, |
| { |
| "epoch": 2.5619047619047617, |
| "grad_norm": 0.930878221988678, |
| "learning_rate": 4.90909090909091e-06, |
| "loss": 0.09620843827724457, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10099, |
| "step": 538, |
| "tokens/total": 2203648, |
| "tokens/train_per_sec_per_gpu": 17.37, |
| "tokens/trainable": 157750 |
| }, |
| { |
| "epoch": 2.5666666666666664, |
| "grad_norm": 0.9390467405319214, |
| "learning_rate": 4.899521531100479e-06, |
| "loss": 0.1333330124616623, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.14263, |
| "step": 539, |
| "tokens/total": 2207744, |
| "tokens/train_per_sec_per_gpu": 20.2, |
| "tokens/trainable": 158105 |
| }, |
| { |
| "epoch": 2.571428571428571, |
| "grad_norm": 0.8828926682472229, |
| "learning_rate": 4.889952153110048e-06, |
| "loss": 0.14065320789813995, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15103, |
| "step": 540, |
| "tokens/total": 2211840, |
| "tokens/train_per_sec_per_gpu": 12.07, |
| "tokens/trainable": 158385 |
| }, |
| { |
| "epoch": 2.576190476190476, |
| "grad_norm": 0.6615021824836731, |
| "learning_rate": 4.880382775119618e-06, |
| "loss": 0.06263934075832367, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06464, |
| "step": 541, |
| "tokens/total": 2215936, |
| "tokens/train_per_sec_per_gpu": 15.48, |
| "tokens/trainable": 158690 |
| }, |
| { |
| "epoch": 2.580952380952381, |
| "grad_norm": 0.7471950650215149, |
| "learning_rate": 4.870813397129187e-06, |
| "loss": 0.06098947301506996, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.06289, |
| "step": 542, |
| "tokens/total": 2220032, |
| "tokens/train_per_sec_per_gpu": 16.24, |
| "tokens/trainable": 158963 |
| }, |
| { |
| "epoch": 2.585714285714286, |
| "grad_norm": 0.856507420539856, |
| "learning_rate": 4.8612440191387565e-06, |
| "loss": 0.06497907638549805, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06714, |
| "step": 543, |
| "tokens/total": 2224128, |
| "tokens/train_per_sec_per_gpu": 10.19, |
| "tokens/trainable": 159236 |
| }, |
| { |
| "epoch": 2.5904761904761906, |
| "grad_norm": 0.7332066893577576, |
| "learning_rate": 4.851674641148326e-06, |
| "loss": 0.08813416957855225, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09213, |
| "step": 544, |
| "tokens/total": 2228224, |
| "tokens/train_per_sec_per_gpu": 9.67, |
| "tokens/trainable": 159454 |
| }, |
| { |
| "epoch": 2.5952380952380953, |
| "grad_norm": 0.8077099323272705, |
| "learning_rate": 4.842105263157895e-06, |
| "loss": 0.07312612235546112, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07587, |
| "step": 545, |
| "tokens/total": 2232320, |
| "tokens/train_per_sec_per_gpu": 10.61, |
| "tokens/trainable": 159676 |
| }, |
| { |
| "epoch": 2.6, |
| "grad_norm": 0.6955282092094421, |
| "learning_rate": 4.832535885167465e-06, |
| "loss": 0.09943203628063202, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10454, |
| "step": 546, |
| "tokens/total": 2236416, |
| "tokens/train_per_sec_per_gpu": 18.31, |
| "tokens/trainable": 160051 |
| }, |
| { |
| "epoch": 2.604761904761905, |
| "grad_norm": 0.835927426815033, |
| "learning_rate": 4.822966507177034e-06, |
| "loss": 0.06828266382217407, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07067, |
| "step": 547, |
| "tokens/total": 2240512, |
| "tokens/train_per_sec_per_gpu": 12.04, |
| "tokens/trainable": 160324 |
| }, |
| { |
| "epoch": 2.6095238095238096, |
| "grad_norm": 0.5681146383285522, |
| "learning_rate": 4.8133971291866035e-06, |
| "loss": 0.05898406356573105, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06076, |
| "step": 548, |
| "tokens/total": 2244608, |
| "tokens/train_per_sec_per_gpu": 12.42, |
| "tokens/trainable": 160605 |
| }, |
| { |
| "epoch": 2.6142857142857143, |
| "grad_norm": 0.584089457988739, |
| "learning_rate": 4.803827751196173e-06, |
| "loss": 0.05229197070002556, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.05368, |
| "step": 549, |
| "tokens/total": 2248704, |
| "tokens/train_per_sec_per_gpu": 14.28, |
| "tokens/trainable": 160847 |
| }, |
| { |
| "epoch": 2.619047619047619, |
| "grad_norm": 0.5604119300842285, |
| "learning_rate": 4.794258373205741e-06, |
| "loss": 0.06550096720457077, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06769, |
| "step": 550, |
| "tokens/total": 2252800, |
| "tokens/train_per_sec_per_gpu": 13.77, |
| "tokens/trainable": 161110 |
| }, |
| { |
| "epoch": 2.623809523809524, |
| "grad_norm": 0.6032201051712036, |
| "learning_rate": 4.784688995215312e-06, |
| "loss": 0.08309999108314514, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08665, |
| "step": 551, |
| "tokens/total": 2256896, |
| "tokens/train_per_sec_per_gpu": 9.8, |
| "tokens/trainable": 161423 |
| }, |
| { |
| "epoch": 2.6285714285714286, |
| "grad_norm": 0.5156391859054565, |
| "learning_rate": 4.775119617224881e-06, |
| "loss": 0.05599772557616234, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0576, |
| "step": 552, |
| "tokens/total": 2260992, |
| "tokens/train_per_sec_per_gpu": 17.55, |
| "tokens/trainable": 161784 |
| }, |
| { |
| "epoch": 2.6333333333333333, |
| "grad_norm": 0.8425244092941284, |
| "learning_rate": 4.76555023923445e-06, |
| "loss": 0.13661256432533264, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14638, |
| "step": 553, |
| "tokens/total": 2265088, |
| "tokens/train_per_sec_per_gpu": 15.11, |
| "tokens/trainable": 162103 |
| }, |
| { |
| "epoch": 2.638095238095238, |
| "grad_norm": 0.6089537143707275, |
| "learning_rate": 4.75598086124402e-06, |
| "loss": 0.05394409969449043, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05543, |
| "step": 554, |
| "tokens/total": 2269184, |
| "tokens/train_per_sec_per_gpu": 13.95, |
| "tokens/trainable": 162339 |
| }, |
| { |
| "epoch": 2.642857142857143, |
| "grad_norm": 0.7857866883277893, |
| "learning_rate": 4.746411483253589e-06, |
| "loss": 0.08736590296030045, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0913, |
| "step": 555, |
| "tokens/total": 2273280, |
| "tokens/train_per_sec_per_gpu": 11.75, |
| "tokens/trainable": 162604 |
| }, |
| { |
| "epoch": 2.6476190476190475, |
| "grad_norm": 0.5566453337669373, |
| "learning_rate": 4.736842105263158e-06, |
| "loss": 0.04392281174659729, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0449, |
| "step": 556, |
| "tokens/total": 2277376, |
| "tokens/train_per_sec_per_gpu": 10.74, |
| "tokens/trainable": 162853 |
| }, |
| { |
| "epoch": 2.6523809523809523, |
| "grad_norm": 0.7361345291137695, |
| "learning_rate": 4.727272727272728e-06, |
| "loss": 0.08509525656700134, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08882, |
| "step": 557, |
| "tokens/total": 2281472, |
| "tokens/train_per_sec_per_gpu": 11.46, |
| "tokens/trainable": 163179 |
| }, |
| { |
| "epoch": 2.657142857142857, |
| "grad_norm": 0.740824818611145, |
| "learning_rate": 4.717703349282297e-06, |
| "loss": 0.07808547466993332, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08122, |
| "step": 558, |
| "tokens/total": 2285568, |
| "tokens/train_per_sec_per_gpu": 20.6, |
| "tokens/trainable": 163529 |
| }, |
| { |
| "epoch": 2.6619047619047618, |
| "grad_norm": 0.5467495918273926, |
| "learning_rate": 4.708133971291867e-06, |
| "loss": 0.035560429096221924, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0362, |
| "step": 559, |
| "tokens/total": 2289664, |
| "tokens/train_per_sec_per_gpu": 10.29, |
| "tokens/trainable": 163778 |
| }, |
| { |
| "epoch": 2.6666666666666665, |
| "grad_norm": 0.9296587705612183, |
| "learning_rate": 4.698564593301435e-06, |
| "loss": 0.09681244194507599, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10165, |
| "step": 560, |
| "tokens/total": 2293760, |
| "tokens/train_per_sec_per_gpu": 10.78, |
| "tokens/trainable": 164042 |
| }, |
| { |
| "epoch": 2.6714285714285713, |
| "grad_norm": 0.6016287207603455, |
| "learning_rate": 4.6889952153110055e-06, |
| "loss": 0.044333577156066895, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04533, |
| "step": 561, |
| "tokens/total": 2297856, |
| "tokens/train_per_sec_per_gpu": 9.94, |
| "tokens/trainable": 164297 |
| }, |
| { |
| "epoch": 2.6761904761904765, |
| "grad_norm": 0.7982633709907532, |
| "learning_rate": 4.679425837320575e-06, |
| "loss": 0.10645254701375961, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11233, |
| "step": 562, |
| "tokens/total": 2301952, |
| "tokens/train_per_sec_per_gpu": 8.47, |
| "tokens/trainable": 164585 |
| }, |
| { |
| "epoch": 2.680952380952381, |
| "grad_norm": 1.001904010772705, |
| "learning_rate": 4.6698564593301435e-06, |
| "loss": 0.11368199437856674, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.1204, |
| "step": 563, |
| "tokens/total": 2306048, |
| "tokens/train_per_sec_per_gpu": 19.83, |
| "tokens/trainable": 164966 |
| }, |
| { |
| "epoch": 2.685714285714286, |
| "grad_norm": 0.5173723101615906, |
| "learning_rate": 4.660287081339714e-06, |
| "loss": 0.05308505892753601, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05452, |
| "step": 564, |
| "tokens/total": 2310144, |
| "tokens/train_per_sec_per_gpu": 11.21, |
| "tokens/trainable": 165258 |
| }, |
| { |
| "epoch": 2.6904761904761907, |
| "grad_norm": 0.70432448387146, |
| "learning_rate": 4.650717703349283e-06, |
| "loss": 0.10319502651691437, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10871, |
| "step": 565, |
| "tokens/total": 2314240, |
| "tokens/train_per_sec_per_gpu": 17.41, |
| "tokens/trainable": 165578 |
| }, |
| { |
| "epoch": 2.6952380952380954, |
| "grad_norm": 1.015570878982544, |
| "learning_rate": 4.641148325358852e-06, |
| "loss": 0.15254569053649902, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1648, |
| "step": 566, |
| "tokens/total": 2318336, |
| "tokens/train_per_sec_per_gpu": 13.67, |
| "tokens/trainable": 165848 |
| }, |
| { |
| "epoch": 2.7, |
| "grad_norm": 0.7950270175933838, |
| "learning_rate": 4.631578947368421e-06, |
| "loss": 0.10250667482614517, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10794, |
| "step": 567, |
| "tokens/total": 2322432, |
| "tokens/train_per_sec_per_gpu": 11.14, |
| "tokens/trainable": 166144 |
| }, |
| { |
| "epoch": 2.704761904761905, |
| "grad_norm": 1.0225125551223755, |
| "learning_rate": 4.622009569377991e-06, |
| "loss": 0.0614117756485939, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06334, |
| "step": 568, |
| "tokens/total": 2326528, |
| "tokens/train_per_sec_per_gpu": 13.63, |
| "tokens/trainable": 166421 |
| }, |
| { |
| "epoch": 2.7095238095238097, |
| "grad_norm": 0.693941593170166, |
| "learning_rate": 4.61244019138756e-06, |
| "loss": 0.0860099121928215, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08982, |
| "step": 569, |
| "tokens/total": 2330624, |
| "tokens/train_per_sec_per_gpu": 12.06, |
| "tokens/trainable": 166693 |
| }, |
| { |
| "epoch": 2.7142857142857144, |
| "grad_norm": 1.0889708995819092, |
| "learning_rate": 4.602870813397129e-06, |
| "loss": 0.09744396805763245, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10235, |
| "step": 570, |
| "tokens/total": 2334720, |
| "tokens/train_per_sec_per_gpu": 9.0, |
| "tokens/trainable": 167017 |
| }, |
| { |
| "epoch": 2.719047619047619, |
| "grad_norm": 0.6773040294647217, |
| "learning_rate": 4.593301435406699e-06, |
| "loss": 0.07057362049818039, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07312, |
| "step": 571, |
| "tokens/total": 2338816, |
| "tokens/train_per_sec_per_gpu": 11.76, |
| "tokens/trainable": 167316 |
| }, |
| { |
| "epoch": 2.723809523809524, |
| "grad_norm": 0.5719726085662842, |
| "learning_rate": 4.583732057416269e-06, |
| "loss": 0.10331287980079651, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10884, |
| "step": 572, |
| "tokens/total": 2342912, |
| "tokens/train_per_sec_per_gpu": 20.31, |
| "tokens/trainable": 167705 |
| }, |
| { |
| "epoch": 2.7285714285714286, |
| "grad_norm": 0.8189126253128052, |
| "learning_rate": 4.574162679425837e-06, |
| "loss": 0.0910937637090683, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09537, |
| "step": 573, |
| "tokens/total": 2347008, |
| "tokens/train_per_sec_per_gpu": 12.8, |
| "tokens/trainable": 167965 |
| }, |
| { |
| "epoch": 2.7333333333333334, |
| "grad_norm": 0.9918308854103088, |
| "learning_rate": 4.564593301435407e-06, |
| "loss": 0.09861594438552856, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10364, |
| "step": 574, |
| "tokens/total": 2351104, |
| "tokens/train_per_sec_per_gpu": 10.17, |
| "tokens/trainable": 168213 |
| }, |
| { |
| "epoch": 2.738095238095238, |
| "grad_norm": 0.5878109931945801, |
| "learning_rate": 4.555023923444977e-06, |
| "loss": 0.07064704596996307, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0732, |
| "step": 575, |
| "tokens/total": 2355200, |
| "tokens/train_per_sec_per_gpu": 12.66, |
| "tokens/trainable": 168475 |
| }, |
| { |
| "epoch": 2.742857142857143, |
| "grad_norm": 0.5074825882911682, |
| "learning_rate": 4.5454545454545455e-06, |
| "loss": 0.026333343237638474, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02668, |
| "step": 576, |
| "tokens/total": 2359296, |
| "tokens/train_per_sec_per_gpu": 10.61, |
| "tokens/trainable": 168718 |
| }, |
| { |
| "epoch": 2.7476190476190476, |
| "grad_norm": 0.9708786606788635, |
| "learning_rate": 4.535885167464115e-06, |
| "loss": 0.126325786113739, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13465, |
| "step": 577, |
| "tokens/total": 2363392, |
| "tokens/train_per_sec_per_gpu": 21.11, |
| "tokens/trainable": 169106 |
| }, |
| { |
| "epoch": 2.7523809523809524, |
| "grad_norm": 0.6538110375404358, |
| "learning_rate": 4.526315789473685e-06, |
| "loss": 0.07794515043497086, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08106, |
| "step": 578, |
| "tokens/total": 2367488, |
| "tokens/train_per_sec_per_gpu": 11.35, |
| "tokens/trainable": 169363 |
| }, |
| { |
| "epoch": 2.757142857142857, |
| "grad_norm": 0.5316607356071472, |
| "learning_rate": 4.516746411483254e-06, |
| "loss": 0.039749983698129654, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04055, |
| "step": 579, |
| "tokens/total": 2371584, |
| "tokens/train_per_sec_per_gpu": 15.3, |
| "tokens/trainable": 169673 |
| }, |
| { |
| "epoch": 2.761904761904762, |
| "grad_norm": 0.8430630564689636, |
| "learning_rate": 4.507177033492823e-06, |
| "loss": 0.11248931288719177, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11906, |
| "step": 580, |
| "tokens/total": 2375680, |
| "tokens/train_per_sec_per_gpu": 17.75, |
| "tokens/trainable": 170058 |
| }, |
| { |
| "epoch": 2.7666666666666666, |
| "grad_norm": 0.6618925333023071, |
| "learning_rate": 4.4976076555023925e-06, |
| "loss": 0.10998591780662537, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11626, |
| "step": 581, |
| "tokens/total": 2379776, |
| "tokens/train_per_sec_per_gpu": 10.85, |
| "tokens/trainable": 170370 |
| }, |
| { |
| "epoch": 2.7714285714285714, |
| "grad_norm": 1.112197995185852, |
| "learning_rate": 4.488038277511962e-06, |
| "loss": 0.17260217666625977, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.18839, |
| "step": 582, |
| "tokens/total": 2383872, |
| "tokens/train_per_sec_per_gpu": 15.45, |
| "tokens/trainable": 170672 |
| }, |
| { |
| "epoch": 2.776190476190476, |
| "grad_norm": 0.9261051416397095, |
| "learning_rate": 4.478468899521531e-06, |
| "loss": 0.06308593600988388, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06512, |
| "step": 583, |
| "tokens/total": 2387968, |
| "tokens/train_per_sec_per_gpu": 9.88, |
| "tokens/trainable": 170922 |
| }, |
| { |
| "epoch": 2.780952380952381, |
| "grad_norm": 0.8806402683258057, |
| "learning_rate": 4.468899521531101e-06, |
| "loss": 0.10380478203296661, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10938, |
| "step": 584, |
| "tokens/total": 2392064, |
| "tokens/train_per_sec_per_gpu": 14.02, |
| "tokens/trainable": 171222 |
| }, |
| { |
| "epoch": 2.7857142857142856, |
| "grad_norm": 0.4549677073955536, |
| "learning_rate": 4.45933014354067e-06, |
| "loss": 0.031093571335077286, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03158, |
| "step": 585, |
| "tokens/total": 2396160, |
| "tokens/train_per_sec_per_gpu": 16.84, |
| "tokens/trainable": 171563 |
| }, |
| { |
| "epoch": 2.7904761904761903, |
| "grad_norm": 1.1855583190917969, |
| "learning_rate": 4.449760765550239e-06, |
| "loss": 0.09628066420555115, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10107, |
| "step": 586, |
| "tokens/total": 2400256, |
| "tokens/train_per_sec_per_gpu": 14.16, |
| "tokens/trainable": 171843 |
| }, |
| { |
| "epoch": 2.795238095238095, |
| "grad_norm": 0.8998063206672668, |
| "learning_rate": 4.440191387559809e-06, |
| "loss": 0.100868821144104, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10613, |
| "step": 587, |
| "tokens/total": 2404352, |
| "tokens/train_per_sec_per_gpu": 8.07, |
| "tokens/trainable": 172142 |
| }, |
| { |
| "epoch": 2.8, |
| "grad_norm": 0.5062983632087708, |
| "learning_rate": 4.430622009569378e-06, |
| "loss": 0.05945008993148804, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06125, |
| "step": 588, |
| "tokens/total": 2408448, |
| "tokens/train_per_sec_per_gpu": 12.06, |
| "tokens/trainable": 172464 |
| }, |
| { |
| "epoch": 2.8047619047619046, |
| "grad_norm": 0.6035679578781128, |
| "learning_rate": 4.4210526315789476e-06, |
| "loss": 0.0527825802564621, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0542, |
| "step": 589, |
| "tokens/total": 2412544, |
| "tokens/train_per_sec_per_gpu": 13.63, |
| "tokens/trainable": 172734 |
| }, |
| { |
| "epoch": 2.8095238095238093, |
| "grad_norm": 0.6835501790046692, |
| "learning_rate": 4.411483253588517e-06, |
| "loss": 0.06511889398097992, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06729, |
| "step": 590, |
| "tokens/total": 2416640, |
| "tokens/train_per_sec_per_gpu": 13.47, |
| "tokens/trainable": 172989 |
| }, |
| { |
| "epoch": 2.814285714285714, |
| "grad_norm": 0.8977090716362, |
| "learning_rate": 4.401913875598086e-06, |
| "loss": 0.07419802248477936, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07702, |
| "step": 591, |
| "tokens/total": 2420736, |
| "tokens/train_per_sec_per_gpu": 8.11, |
| "tokens/trainable": 173233 |
| }, |
| { |
| "epoch": 2.819047619047619, |
| "grad_norm": 0.522295355796814, |
| "learning_rate": 4.392344497607656e-06, |
| "loss": 0.05964972823858261, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06146, |
| "step": 592, |
| "tokens/total": 2424832, |
| "tokens/train_per_sec_per_gpu": 11.41, |
| "tokens/trainable": 173511 |
| }, |
| { |
| "epoch": 2.8238095238095235, |
| "grad_norm": 0.8135259747505188, |
| "learning_rate": 4.382775119617225e-06, |
| "loss": 0.0776074081659317, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0807, |
| "step": 593, |
| "tokens/total": 2428928, |
| "tokens/train_per_sec_per_gpu": 9.37, |
| "tokens/trainable": 173915 |
| }, |
| { |
| "epoch": 2.8285714285714287, |
| "grad_norm": 0.773694634437561, |
| "learning_rate": 4.3732057416267945e-06, |
| "loss": 0.09940958023071289, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10452, |
| "step": 594, |
| "tokens/total": 2433024, |
| "tokens/train_per_sec_per_gpu": 15.03, |
| "tokens/trainable": 174184 |
| }, |
| { |
| "epoch": 2.8333333333333335, |
| "grad_norm": 0.8801134824752808, |
| "learning_rate": 4.363636363636364e-06, |
| "loss": 0.1124865710735321, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11906, |
| "step": 595, |
| "tokens/total": 2437120, |
| "tokens/train_per_sec_per_gpu": 16.28, |
| "tokens/trainable": 174522 |
| }, |
| { |
| "epoch": 2.8380952380952382, |
| "grad_norm": 0.48666656017303467, |
| "learning_rate": 4.354066985645933e-06, |
| "loss": 0.03523673117160797, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.03586, |
| "step": 596, |
| "tokens/total": 2441216, |
| "tokens/train_per_sec_per_gpu": 14.43, |
| "tokens/trainable": 174842 |
| }, |
| { |
| "epoch": 2.842857142857143, |
| "grad_norm": 0.5700379014015198, |
| "learning_rate": 4.344497607655503e-06, |
| "loss": 0.059006646275520325, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06078, |
| "step": 597, |
| "tokens/total": 2445312, |
| "tokens/train_per_sec_per_gpu": 19.05, |
| "tokens/trainable": 175180 |
| }, |
| { |
| "epoch": 2.8476190476190477, |
| "grad_norm": 0.7380210757255554, |
| "learning_rate": 4.334928229665072e-06, |
| "loss": 0.12105896323919296, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12869, |
| "step": 598, |
| "tokens/total": 2449408, |
| "tokens/train_per_sec_per_gpu": 16.09, |
| "tokens/trainable": 175549 |
| }, |
| { |
| "epoch": 2.8523809523809525, |
| "grad_norm": 0.7186801433563232, |
| "learning_rate": 4.3253588516746415e-06, |
| "loss": 0.05410782992839813, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0556, |
| "step": 599, |
| "tokens/total": 2453504, |
| "tokens/train_per_sec_per_gpu": 8.99, |
| "tokens/trainable": 175821 |
| }, |
| { |
| "epoch": 2.857142857142857, |
| "grad_norm": 0.8128309845924377, |
| "learning_rate": 4.315789473684211e-06, |
| "loss": 0.062483049929142, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06448, |
| "step": 600, |
| "tokens/total": 2457600, |
| "tokens/train_per_sec_per_gpu": 15.0, |
| "tokens/trainable": 176090 |
| }, |
| { |
| "epoch": 2.861904761904762, |
| "grad_norm": 0.6894558072090149, |
| "learning_rate": 4.30622009569378e-06, |
| "loss": 0.060672447085380554, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06255, |
| "step": 601, |
| "tokens/total": 2461696, |
| "tokens/train_per_sec_per_gpu": 17.7, |
| "tokens/trainable": 176399 |
| }, |
| { |
| "epoch": 2.8666666666666667, |
| "grad_norm": 0.54118812084198, |
| "learning_rate": 4.29665071770335e-06, |
| "loss": 0.050571709871292114, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05187, |
| "step": 602, |
| "tokens/total": 2465792, |
| "tokens/train_per_sec_per_gpu": 12.01, |
| "tokens/trainable": 176665 |
| }, |
| { |
| "epoch": 2.8714285714285714, |
| "grad_norm": 0.7929149866104126, |
| "learning_rate": 4.287081339712919e-06, |
| "loss": 0.0880894660949707, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09209, |
| "step": 603, |
| "tokens/total": 2469888, |
| "tokens/train_per_sec_per_gpu": 15.68, |
| "tokens/trainable": 176985 |
| }, |
| { |
| "epoch": 2.876190476190476, |
| "grad_norm": 0.7643014192581177, |
| "learning_rate": 4.277511961722488e-06, |
| "loss": 0.06582175940275192, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06804, |
| "step": 604, |
| "tokens/total": 2473984, |
| "tokens/train_per_sec_per_gpu": 15.27, |
| "tokens/trainable": 177325 |
| }, |
| { |
| "epoch": 2.880952380952381, |
| "grad_norm": 0.5427528023719788, |
| "learning_rate": 4.267942583732058e-06, |
| "loss": 0.06989608705043793, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0724, |
| "step": 605, |
| "tokens/total": 2478080, |
| "tokens/train_per_sec_per_gpu": 16.05, |
| "tokens/trainable": 177646 |
| }, |
| { |
| "epoch": 2.8857142857142857, |
| "grad_norm": 0.698477566242218, |
| "learning_rate": 4.258373205741627e-06, |
| "loss": 0.08636561036109924, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0902, |
| "step": 606, |
| "tokens/total": 2482176, |
| "tokens/train_per_sec_per_gpu": 11.15, |
| "tokens/trainable": 177970 |
| }, |
| { |
| "epoch": 2.8904761904761904, |
| "grad_norm": 1.405834674835205, |
| "learning_rate": 4.2488038277511966e-06, |
| "loss": 0.09993740916252136, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1051, |
| "step": 607, |
| "tokens/total": 2486272, |
| "tokens/train_per_sec_per_gpu": 13.81, |
| "tokens/trainable": 178230 |
| }, |
| { |
| "epoch": 2.895238095238095, |
| "grad_norm": 1.0274302959442139, |
| "learning_rate": 4.239234449760766e-06, |
| "loss": 0.10263222455978394, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10808, |
| "step": 608, |
| "tokens/total": 2490368, |
| "tokens/train_per_sec_per_gpu": 13.62, |
| "tokens/trainable": 178484 |
| }, |
| { |
| "epoch": 2.9, |
| "grad_norm": 1.4119830131530762, |
| "learning_rate": 4.229665071770335e-06, |
| "loss": 0.11618393659591675, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1232, |
| "step": 609, |
| "tokens/total": 2494464, |
| "tokens/train_per_sec_per_gpu": 11.03, |
| "tokens/trainable": 178759 |
| }, |
| { |
| "epoch": 2.9047619047619047, |
| "grad_norm": 0.9467605352401733, |
| "learning_rate": 4.220095693779905e-06, |
| "loss": 0.08310011774301529, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08665, |
| "step": 610, |
| "tokens/total": 2498560, |
| "tokens/train_per_sec_per_gpu": 12.83, |
| "tokens/trainable": 179007 |
| }, |
| { |
| "epoch": 2.9095238095238094, |
| "grad_norm": 0.8503310680389404, |
| "learning_rate": 4.210526315789474e-06, |
| "loss": 0.14148107171058655, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15198, |
| "step": 611, |
| "tokens/total": 2502656, |
| "tokens/train_per_sec_per_gpu": 14.39, |
| "tokens/trainable": 179326 |
| }, |
| { |
| "epoch": 2.914285714285714, |
| "grad_norm": 0.789787232875824, |
| "learning_rate": 4.2009569377990435e-06, |
| "loss": 0.04962185025215149, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05087, |
| "step": 612, |
| "tokens/total": 2506752, |
| "tokens/train_per_sec_per_gpu": 12.14, |
| "tokens/trainable": 179559 |
| }, |
| { |
| "epoch": 2.919047619047619, |
| "grad_norm": 0.39260679483413696, |
| "learning_rate": 4.191387559808613e-06, |
| "loss": 0.031790271401405334, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0323, |
| "step": 613, |
| "tokens/total": 2510848, |
| "tokens/train_per_sec_per_gpu": 18.35, |
| "tokens/trainable": 179872 |
| }, |
| { |
| "epoch": 2.923809523809524, |
| "grad_norm": 0.9382997155189514, |
| "learning_rate": 4.181818181818182e-06, |
| "loss": 0.12968839704990387, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13847, |
| "step": 614, |
| "tokens/total": 2514944, |
| "tokens/train_per_sec_per_gpu": 15.59, |
| "tokens/trainable": 180147 |
| }, |
| { |
| "epoch": 2.928571428571429, |
| "grad_norm": 8.853306770324707, |
| "learning_rate": 4.172248803827752e-06, |
| "loss": 0.10713415592908859, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11308, |
| "step": 615, |
| "tokens/total": 2519040, |
| "tokens/train_per_sec_per_gpu": 11.89, |
| "tokens/trainable": 180419 |
| }, |
| { |
| "epoch": 2.9333333333333336, |
| "grad_norm": 0.6513184309005737, |
| "learning_rate": 4.162679425837321e-06, |
| "loss": 0.07565292716026306, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07859, |
| "step": 616, |
| "tokens/total": 2523136, |
| "tokens/train_per_sec_per_gpu": 17.6, |
| "tokens/trainable": 180756 |
| }, |
| { |
| "epoch": 2.9380952380952383, |
| "grad_norm": 0.6604302525520325, |
| "learning_rate": 4.1531100478468904e-06, |
| "loss": 0.04975314065814018, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05101, |
| "step": 617, |
| "tokens/total": 2527232, |
| "tokens/train_per_sec_per_gpu": 13.61, |
| "tokens/trainable": 181020 |
| }, |
| { |
| "epoch": 2.942857142857143, |
| "grad_norm": 0.7588087320327759, |
| "learning_rate": 4.14354066985646e-06, |
| "loss": 0.06836371123790741, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07075, |
| "step": 618, |
| "tokens/total": 2531328, |
| "tokens/train_per_sec_per_gpu": 10.98, |
| "tokens/trainable": 181242 |
| }, |
| { |
| "epoch": 2.947619047619048, |
| "grad_norm": 0.6721870303153992, |
| "learning_rate": 4.133971291866029e-06, |
| "loss": 0.10186492651700974, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.10723, |
| "step": 619, |
| "tokens/total": 2535424, |
| "tokens/train_per_sec_per_gpu": 17.53, |
| "tokens/trainable": 181528 |
| }, |
| { |
| "epoch": 2.9523809523809526, |
| "grad_norm": 0.674967348575592, |
| "learning_rate": 4.124401913875599e-06, |
| "loss": 0.05646386370062828, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05809, |
| "step": 620, |
| "tokens/total": 2539520, |
| "tokens/train_per_sec_per_gpu": 10.12, |
| "tokens/trainable": 181783 |
| }, |
| { |
| "epoch": 2.9571428571428573, |
| "grad_norm": 1.0123529434204102, |
| "learning_rate": 4.114832535885168e-06, |
| "loss": 0.085752934217453, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08954, |
| "step": 621, |
| "tokens/total": 2543616, |
| "tokens/train_per_sec_per_gpu": 12.18, |
| "tokens/trainable": 182014 |
| }, |
| { |
| "epoch": 2.961904761904762, |
| "grad_norm": 0.9926598072052002, |
| "learning_rate": 4.105263157894737e-06, |
| "loss": 0.1454455703496933, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.15655, |
| "step": 622, |
| "tokens/total": 2547712, |
| "tokens/train_per_sec_per_gpu": 9.38, |
| "tokens/trainable": 182234 |
| }, |
| { |
| "epoch": 2.966666666666667, |
| "grad_norm": 0.7257160544395447, |
| "learning_rate": 4.095693779904307e-06, |
| "loss": 0.09883573651313782, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.10388, |
| "step": 623, |
| "tokens/total": 2551808, |
| "tokens/train_per_sec_per_gpu": 24.54, |
| "tokens/trainable": 182626 |
| }, |
| { |
| "epoch": 2.9714285714285715, |
| "grad_norm": 1.0070229768753052, |
| "learning_rate": 4.086124401913875e-06, |
| "loss": 0.07697214186191559, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08001, |
| "step": 624, |
| "tokens/total": 2555904, |
| "tokens/train_per_sec_per_gpu": 14.22, |
| "tokens/trainable": 182916 |
| }, |
| { |
| "epoch": 2.9761904761904763, |
| "grad_norm": 0.6566667556762695, |
| "learning_rate": 4.0765550239234456e-06, |
| "loss": 0.07002940773963928, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07254, |
| "step": 625, |
| "tokens/total": 2560000, |
| "tokens/train_per_sec_per_gpu": 10.01, |
| "tokens/trainable": 183194 |
| }, |
| { |
| "epoch": 2.980952380952381, |
| "grad_norm": 0.9295153021812439, |
| "learning_rate": 4.066985645933015e-06, |
| "loss": 0.10709331929683685, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11304, |
| "step": 626, |
| "tokens/total": 2564096, |
| "tokens/train_per_sec_per_gpu": 17.2, |
| "tokens/trainable": 183592 |
| }, |
| { |
| "epoch": 2.9857142857142858, |
| "grad_norm": 0.5641553401947021, |
| "learning_rate": 4.0574162679425835e-06, |
| "loss": 0.06261751800775528, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06462, |
| "step": 627, |
| "tokens/total": 2568192, |
| "tokens/train_per_sec_per_gpu": 18.24, |
| "tokens/trainable": 183885 |
| }, |
| { |
| "epoch": 2.9904761904761905, |
| "grad_norm": 0.3403347134590149, |
| "learning_rate": 4.047846889952154e-06, |
| "loss": 0.04572930932044983, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04679, |
| "step": 628, |
| "tokens/total": 2572288, |
| "tokens/train_per_sec_per_gpu": 20.39, |
| "tokens/trainable": 184308 |
| }, |
| { |
| "epoch": 2.9952380952380953, |
| "grad_norm": 0.5752260088920593, |
| "learning_rate": 4.038277511961723e-06, |
| "loss": 0.052835218608379364, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05426, |
| "step": 629, |
| "tokens/total": 2576384, |
| "tokens/train_per_sec_per_gpu": 17.39, |
| "tokens/trainable": 184586 |
| }, |
| { |
| "epoch": 3.0, |
| "grad_norm": 0.7610874176025391, |
| "learning_rate": 4.0287081339712925e-06, |
| "loss": 0.06551462411880493, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06771, |
| "step": 630, |
| "tokens/total": 2580480, |
| "tokens/train_per_sec_per_gpu": 14.51, |
| "tokens/trainable": 184865 |
| }, |
| { |
| "epoch": 3.0, |
| "eval_loss": 0.083167664706707, |
| "eval_ppl": 1.08672, |
| "eval_runtime": 10.3423, |
| "eval_samples_per_second": 8.605, |
| "eval_steps_per_second": 2.224, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 35.58, |
| "memory/max_allocated (GiB)": 35.58, |
| "step": 630 |
| }, |
| { |
| "epoch": 3.0047619047619047, |
| "grad_norm": 1.1097360849380493, |
| "learning_rate": 4.019138755980861e-06, |
| "loss": 0.027190648019313812, |
| "memory/device_reserved (GiB)": 39.15, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02756, |
| "step": 631, |
| "tokens/total": 2584576, |
| "tokens/train_per_sec_per_gpu": 9.62, |
| "tokens/trainable": 185139 |
| }, |
| { |
| "epoch": 3.0095238095238095, |
| "grad_norm": 0.8078517317771912, |
| "learning_rate": 4.009569377990431e-06, |
| "loss": 0.0643414855003357, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06646, |
| "step": 632, |
| "tokens/total": 2588672, |
| "tokens/train_per_sec_per_gpu": 8.58, |
| "tokens/trainable": 185352 |
| }, |
| { |
| "epoch": 3.0142857142857142, |
| "grad_norm": 0.9028353095054626, |
| "learning_rate": 4.000000000000001e-06, |
| "loss": 0.07669511437416077, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07971, |
| "step": 633, |
| "tokens/total": 2592768, |
| "tokens/train_per_sec_per_gpu": 11.52, |
| "tokens/trainable": 185600 |
| }, |
| { |
| "epoch": 3.019047619047619, |
| "grad_norm": 0.9765618443489075, |
| "learning_rate": 3.990430622009569e-06, |
| "loss": 0.10810641199350357, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11417, |
| "step": 634, |
| "tokens/total": 2596864, |
| "tokens/train_per_sec_per_gpu": 13.33, |
| "tokens/trainable": 185961 |
| }, |
| { |
| "epoch": 3.0238095238095237, |
| "grad_norm": 0.4859309196472168, |
| "learning_rate": 3.9808612440191394e-06, |
| "loss": 0.031761132180690765, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03227, |
| "step": 635, |
| "tokens/total": 2600960, |
| "tokens/train_per_sec_per_gpu": 9.99, |
| "tokens/trainable": 186204 |
| }, |
| { |
| "epoch": 3.0285714285714285, |
| "grad_norm": 0.9322464466094971, |
| "learning_rate": 3.971291866028709e-06, |
| "loss": 0.06739023327827454, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06971, |
| "step": 636, |
| "tokens/total": 2605056, |
| "tokens/train_per_sec_per_gpu": 8.54, |
| "tokens/trainable": 186439 |
| }, |
| { |
| "epoch": 3.033333333333333, |
| "grad_norm": 0.6128789782524109, |
| "learning_rate": 3.961722488038277e-06, |
| "loss": 0.053875766694545746, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.05535, |
| "step": 637, |
| "tokens/total": 2609152, |
| "tokens/train_per_sec_per_gpu": 14.72, |
| "tokens/trainable": 186699 |
| }, |
| { |
| "epoch": 3.038095238095238, |
| "grad_norm": 0.4706403315067291, |
| "learning_rate": 3.952153110047848e-06, |
| "loss": 0.04248778894543648, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0434, |
| "step": 638, |
| "tokens/total": 2613248, |
| "tokens/train_per_sec_per_gpu": 10.07, |
| "tokens/trainable": 186927 |
| }, |
| { |
| "epoch": 3.0428571428571427, |
| "grad_norm": 0.4942407011985779, |
| "learning_rate": 3.942583732057417e-06, |
| "loss": 0.07395349442958832, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07676, |
| "step": 639, |
| "tokens/total": 2617344, |
| "tokens/train_per_sec_per_gpu": 11.42, |
| "tokens/trainable": 187212 |
| }, |
| { |
| "epoch": 3.0476190476190474, |
| "grad_norm": 1.0005759000778198, |
| "learning_rate": 3.9330143540669855e-06, |
| "loss": 0.07736619561910629, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08044, |
| "step": 640, |
| "tokens/total": 2621440, |
| "tokens/train_per_sec_per_gpu": 13.24, |
| "tokens/trainable": 187591 |
| }, |
| { |
| "epoch": 3.052380952380952, |
| "grad_norm": 0.6871622204780579, |
| "learning_rate": 3.923444976076555e-06, |
| "loss": 0.08220869302749634, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08568, |
| "step": 641, |
| "tokens/total": 2625536, |
| "tokens/train_per_sec_per_gpu": 15.39, |
| "tokens/trainable": 187971 |
| }, |
| { |
| "epoch": 3.057142857142857, |
| "grad_norm": 0.9228248000144958, |
| "learning_rate": 3.913875598086125e-06, |
| "loss": 0.08459930121898651, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08828, |
| "step": 642, |
| "tokens/total": 2629632, |
| "tokens/train_per_sec_per_gpu": 11.29, |
| "tokens/trainable": 188217 |
| }, |
| { |
| "epoch": 3.0619047619047617, |
| "grad_norm": 0.6304330229759216, |
| "learning_rate": 3.904306220095694e-06, |
| "loss": 0.1100236177444458, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1163, |
| "step": 643, |
| "tokens/total": 2633728, |
| "tokens/train_per_sec_per_gpu": 21.66, |
| "tokens/trainable": 188584 |
| }, |
| { |
| "epoch": 3.066666666666667, |
| "grad_norm": 0.6323570013046265, |
| "learning_rate": 3.894736842105263e-06, |
| "loss": 0.05556914210319519, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05714, |
| "step": 644, |
| "tokens/total": 2637824, |
| "tokens/train_per_sec_per_gpu": 9.29, |
| "tokens/trainable": 188832 |
| }, |
| { |
| "epoch": 3.0714285714285716, |
| "grad_norm": 0.9141075015068054, |
| "learning_rate": 3.885167464114833e-06, |
| "loss": 0.10358056426048279, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10914, |
| "step": 645, |
| "tokens/total": 2641920, |
| "tokens/train_per_sec_per_gpu": 14.01, |
| "tokens/trainable": 189081 |
| }, |
| { |
| "epoch": 3.0761904761904764, |
| "grad_norm": 0.7243455052375793, |
| "learning_rate": 3.875598086124403e-06, |
| "loss": 0.06565085053443909, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06785, |
| "step": 646, |
| "tokens/total": 2646016, |
| "tokens/train_per_sec_per_gpu": 17.37, |
| "tokens/trainable": 189420 |
| }, |
| { |
| "epoch": 3.080952380952381, |
| "grad_norm": 0.6836820840835571, |
| "learning_rate": 3.866028708133971e-06, |
| "loss": 0.045150648802518845, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04619, |
| "step": 647, |
| "tokens/total": 2650112, |
| "tokens/train_per_sec_per_gpu": 9.75, |
| "tokens/trainable": 189742 |
| }, |
| { |
| "epoch": 3.085714285714286, |
| "grad_norm": 0.6994568109512329, |
| "learning_rate": 3.856459330143541e-06, |
| "loss": 0.0705096647143364, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07305, |
| "step": 648, |
| "tokens/total": 2654208, |
| "tokens/train_per_sec_per_gpu": 17.95, |
| "tokens/trainable": 190070 |
| }, |
| { |
| "epoch": 3.0904761904761906, |
| "grad_norm": 0.7164385914802551, |
| "learning_rate": 3.846889952153111e-06, |
| "loss": 0.06414791941642761, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06625, |
| "step": 649, |
| "tokens/total": 2658304, |
| "tokens/train_per_sec_per_gpu": 9.85, |
| "tokens/trainable": 190354 |
| }, |
| { |
| "epoch": 3.0952380952380953, |
| "grad_norm": 0.5413431525230408, |
| "learning_rate": 3.8373205741626794e-06, |
| "loss": 0.07239377498626709, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07508, |
| "step": 650, |
| "tokens/total": 2662400, |
| "tokens/train_per_sec_per_gpu": 20.42, |
| "tokens/trainable": 190728 |
| }, |
| { |
| "epoch": 3.1, |
| "grad_norm": 0.702539324760437, |
| "learning_rate": 3.827751196172249e-06, |
| "loss": 0.08402359485626221, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08765, |
| "step": 651, |
| "tokens/total": 2666496, |
| "tokens/train_per_sec_per_gpu": 13.93, |
| "tokens/trainable": 190997 |
| }, |
| { |
| "epoch": 3.104761904761905, |
| "grad_norm": 0.6277540326118469, |
| "learning_rate": 3.818181818181819e-06, |
| "loss": 0.0585082545876503, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06025, |
| "step": 652, |
| "tokens/total": 2670592, |
| "tokens/train_per_sec_per_gpu": 8.95, |
| "tokens/trainable": 191308 |
| }, |
| { |
| "epoch": 3.1095238095238096, |
| "grad_norm": 0.6413969397544861, |
| "learning_rate": 3.8086124401913876e-06, |
| "loss": 0.07055683434009552, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07311, |
| "step": 653, |
| "tokens/total": 2674688, |
| "tokens/train_per_sec_per_gpu": 13.89, |
| "tokens/trainable": 191563 |
| }, |
| { |
| "epoch": 3.1142857142857143, |
| "grad_norm": 0.29829949140548706, |
| "learning_rate": 3.7990430622009574e-06, |
| "loss": 0.02009260281920433, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0203, |
| "step": 654, |
| "tokens/total": 2678784, |
| "tokens/train_per_sec_per_gpu": 11.66, |
| "tokens/trainable": 191868 |
| }, |
| { |
| "epoch": 3.119047619047619, |
| "grad_norm": 0.8772543668746948, |
| "learning_rate": 3.789473684210527e-06, |
| "loss": 0.10029996931552887, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1055, |
| "step": 655, |
| "tokens/total": 2682880, |
| "tokens/train_per_sec_per_gpu": 13.11, |
| "tokens/trainable": 192150 |
| }, |
| { |
| "epoch": 3.123809523809524, |
| "grad_norm": 0.6865347027778625, |
| "learning_rate": 3.7799043062200958e-06, |
| "loss": 0.06961806863546371, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0721, |
| "step": 656, |
| "tokens/total": 2686976, |
| "tokens/train_per_sec_per_gpu": 19.12, |
| "tokens/trainable": 192485 |
| }, |
| { |
| "epoch": 3.1285714285714286, |
| "grad_norm": 0.6526204943656921, |
| "learning_rate": 3.770334928229665e-06, |
| "loss": 0.07430330663919449, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07713, |
| "step": 657, |
| "tokens/total": 2691072, |
| "tokens/train_per_sec_per_gpu": 10.48, |
| "tokens/trainable": 192713 |
| }, |
| { |
| "epoch": 3.1333333333333333, |
| "grad_norm": 1.0195298194885254, |
| "learning_rate": 3.760765550239235e-06, |
| "loss": 0.06909658014774323, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07154, |
| "step": 658, |
| "tokens/total": 2695168, |
| "tokens/train_per_sec_per_gpu": 9.38, |
| "tokens/trainable": 192928 |
| }, |
| { |
| "epoch": 3.138095238095238, |
| "grad_norm": 0.8037240505218506, |
| "learning_rate": 3.7511961722488043e-06, |
| "loss": 0.13666372001171112, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14644, |
| "step": 659, |
| "tokens/total": 2699264, |
| "tokens/train_per_sec_per_gpu": 17.36, |
| "tokens/trainable": 193333 |
| }, |
| { |
| "epoch": 3.142857142857143, |
| "grad_norm": 0.8780602216720581, |
| "learning_rate": 3.7416267942583733e-06, |
| "loss": 0.08582095801830292, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08961, |
| "step": 660, |
| "tokens/total": 2703360, |
| "tokens/train_per_sec_per_gpu": 12.44, |
| "tokens/trainable": 193584 |
| }, |
| { |
| "epoch": 3.1476190476190475, |
| "grad_norm": 1.0211085081100464, |
| "learning_rate": 3.732057416267943e-06, |
| "loss": 0.07044003903865814, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07298, |
| "step": 661, |
| "tokens/total": 2707456, |
| "tokens/train_per_sec_per_gpu": 10.54, |
| "tokens/trainable": 193797 |
| }, |
| { |
| "epoch": 3.1523809523809523, |
| "grad_norm": 0.662294328212738, |
| "learning_rate": 3.7224880382775125e-06, |
| "loss": 0.05967669188976288, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06149, |
| "step": 662, |
| "tokens/total": 2711552, |
| "tokens/train_per_sec_per_gpu": 10.95, |
| "tokens/trainable": 194075 |
| }, |
| { |
| "epoch": 3.157142857142857, |
| "grad_norm": 1.0162534713745117, |
| "learning_rate": 3.7129186602870815e-06, |
| "loss": 0.12779566645622253, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13632, |
| "step": 663, |
| "tokens/total": 2715648, |
| "tokens/train_per_sec_per_gpu": 15.44, |
| "tokens/trainable": 194415 |
| }, |
| { |
| "epoch": 3.1619047619047618, |
| "grad_norm": 1.1409763097763062, |
| "learning_rate": 3.703349282296651e-06, |
| "loss": 0.1108705997467041, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11725, |
| "step": 664, |
| "tokens/total": 2719744, |
| "tokens/train_per_sec_per_gpu": 12.88, |
| "tokens/trainable": 194736 |
| }, |
| { |
| "epoch": 3.1666666666666665, |
| "grad_norm": 0.47355273365974426, |
| "learning_rate": 3.6937799043062207e-06, |
| "loss": 0.041094761341810226, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04195, |
| "step": 665, |
| "tokens/total": 2723840, |
| "tokens/train_per_sec_per_gpu": 13.33, |
| "tokens/trainable": 194990 |
| }, |
| { |
| "epoch": 3.1714285714285713, |
| "grad_norm": 0.9870181083679199, |
| "learning_rate": 3.6842105263157896e-06, |
| "loss": 0.09811937063932419, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10309, |
| "step": 666, |
| "tokens/total": 2727936, |
| "tokens/train_per_sec_per_gpu": 14.63, |
| "tokens/trainable": 195264 |
| }, |
| { |
| "epoch": 3.176190476190476, |
| "grad_norm": 0.541223406791687, |
| "learning_rate": 3.674641148325359e-06, |
| "loss": 0.022513944655656815, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02277, |
| "step": 667, |
| "tokens/total": 2732032, |
| "tokens/train_per_sec_per_gpu": 8.97, |
| "tokens/trainable": 195481 |
| }, |
| { |
| "epoch": 3.1809523809523808, |
| "grad_norm": 0.7369751334190369, |
| "learning_rate": 3.665071770334929e-06, |
| "loss": 0.04886762052774429, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05008, |
| "step": 668, |
| "tokens/total": 2736128, |
| "tokens/train_per_sec_per_gpu": 10.45, |
| "tokens/trainable": 195703 |
| }, |
| { |
| "epoch": 3.185714285714286, |
| "grad_norm": 0.8254791498184204, |
| "learning_rate": 3.655502392344498e-06, |
| "loss": 0.06476770341396332, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06691, |
| "step": 669, |
| "tokens/total": 2740224, |
| "tokens/train_per_sec_per_gpu": 14.07, |
| "tokens/trainable": 195954 |
| }, |
| { |
| "epoch": 3.1904761904761907, |
| "grad_norm": 0.8415271043777466, |
| "learning_rate": 3.645933014354067e-06, |
| "loss": 0.09028398990631104, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09449, |
| "step": 670, |
| "tokens/total": 2744320, |
| "tokens/train_per_sec_per_gpu": 13.24, |
| "tokens/trainable": 196251 |
| }, |
| { |
| "epoch": 3.1952380952380954, |
| "grad_norm": 0.5479375123977661, |
| "learning_rate": 3.6363636363636366e-06, |
| "loss": 0.027023475617170334, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02739, |
| "step": 671, |
| "tokens/total": 2748416, |
| "tokens/train_per_sec_per_gpu": 11.25, |
| "tokens/trainable": 196540 |
| }, |
| { |
| "epoch": 3.2, |
| "grad_norm": 0.5744169354438782, |
| "learning_rate": 3.6267942583732064e-06, |
| "loss": 0.04903862625360489, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05026, |
| "step": 672, |
| "tokens/total": 2752512, |
| "tokens/train_per_sec_per_gpu": 13.45, |
| "tokens/trainable": 196903 |
| }, |
| { |
| "epoch": 3.204761904761905, |
| "grad_norm": 1.511560082435608, |
| "learning_rate": 3.6172248803827754e-06, |
| "loss": 0.0802629217505455, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08357, |
| "step": 673, |
| "tokens/total": 2756608, |
| "tokens/train_per_sec_per_gpu": 9.91, |
| "tokens/trainable": 197165 |
| }, |
| { |
| "epoch": 3.2095238095238097, |
| "grad_norm": 0.4563305974006653, |
| "learning_rate": 3.6076555023923447e-06, |
| "loss": 0.041035354137420654, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04189, |
| "step": 674, |
| "tokens/total": 2760704, |
| "tokens/train_per_sec_per_gpu": 11.78, |
| "tokens/trainable": 197468 |
| }, |
| { |
| "epoch": 3.2142857142857144, |
| "grad_norm": 0.5280977487564087, |
| "learning_rate": 3.5980861244019146e-06, |
| "loss": 0.06585241109132767, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.06807, |
| "step": 675, |
| "tokens/total": 2764800, |
| "tokens/train_per_sec_per_gpu": 17.67, |
| "tokens/trainable": 197805 |
| }, |
| { |
| "epoch": 3.219047619047619, |
| "grad_norm": 0.7410734295845032, |
| "learning_rate": 3.5885167464114835e-06, |
| "loss": 0.06726351380348206, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06958, |
| "step": 676, |
| "tokens/total": 2768896, |
| "tokens/train_per_sec_per_gpu": 16.25, |
| "tokens/trainable": 198106 |
| }, |
| { |
| "epoch": 3.223809523809524, |
| "grad_norm": 0.5525427460670471, |
| "learning_rate": 3.578947368421053e-06, |
| "loss": 0.06145021691918373, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06338, |
| "step": 677, |
| "tokens/total": 2772992, |
| "tokens/train_per_sec_per_gpu": 12.63, |
| "tokens/trainable": 198359 |
| }, |
| { |
| "epoch": 3.2285714285714286, |
| "grad_norm": 0.5915558934211731, |
| "learning_rate": 3.5693779904306223e-06, |
| "loss": 0.0412653312087059, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.04213, |
| "step": 678, |
| "tokens/total": 2777088, |
| "tokens/train_per_sec_per_gpu": 13.21, |
| "tokens/trainable": 198670 |
| }, |
| { |
| "epoch": 3.2333333333333334, |
| "grad_norm": 0.5353713631629944, |
| "learning_rate": 3.5598086124401917e-06, |
| "loss": 0.07545843720436096, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07838, |
| "step": 679, |
| "tokens/total": 2781184, |
| "tokens/train_per_sec_per_gpu": 17.55, |
| "tokens/trainable": 199027 |
| }, |
| { |
| "epoch": 3.238095238095238, |
| "grad_norm": 0.7455602288246155, |
| "learning_rate": 3.550239234449761e-06, |
| "loss": 0.08536748588085175, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08912, |
| "step": 680, |
| "tokens/total": 2785280, |
| "tokens/train_per_sec_per_gpu": 14.58, |
| "tokens/trainable": 199292 |
| }, |
| { |
| "epoch": 3.242857142857143, |
| "grad_norm": 0.4214642345905304, |
| "learning_rate": 3.5406698564593305e-06, |
| "loss": 0.028588632121682167, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.029, |
| "step": 681, |
| "tokens/total": 2789376, |
| "tokens/train_per_sec_per_gpu": 10.7, |
| "tokens/trainable": 199570 |
| }, |
| { |
| "epoch": 3.2476190476190476, |
| "grad_norm": 1.102713704109192, |
| "learning_rate": 3.5311004784688994e-06, |
| "loss": 0.07013420760631561, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07265, |
| "step": 682, |
| "tokens/total": 2793472, |
| "tokens/train_per_sec_per_gpu": 13.58, |
| "tokens/trainable": 199877 |
| }, |
| { |
| "epoch": 3.2523809523809524, |
| "grad_norm": 0.9483047127723694, |
| "learning_rate": 3.5215311004784692e-06, |
| "loss": 0.08880291879177094, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09287, |
| "step": 683, |
| "tokens/total": 2797568, |
| "tokens/train_per_sec_per_gpu": 10.23, |
| "tokens/trainable": 200135 |
| }, |
| { |
| "epoch": 3.257142857142857, |
| "grad_norm": 0.7713571786880493, |
| "learning_rate": 3.5119617224880386e-06, |
| "loss": 0.06746732443571091, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0698, |
| "step": 684, |
| "tokens/total": 2801664, |
| "tokens/train_per_sec_per_gpu": 16.56, |
| "tokens/trainable": 200409 |
| }, |
| { |
| "epoch": 3.261904761904762, |
| "grad_norm": 0.7084768414497375, |
| "learning_rate": 3.5023923444976076e-06, |
| "loss": 0.07177010923624039, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07441, |
| "step": 685, |
| "tokens/total": 2805760, |
| "tokens/train_per_sec_per_gpu": 19.25, |
| "tokens/trainable": 200809 |
| }, |
| { |
| "epoch": 3.2666666666666666, |
| "grad_norm": 1.0349971055984497, |
| "learning_rate": 3.4928229665071774e-06, |
| "loss": 0.08891840279102325, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09299, |
| "step": 686, |
| "tokens/total": 2809856, |
| "tokens/train_per_sec_per_gpu": 14.04, |
| "tokens/trainable": 201052 |
| }, |
| { |
| "epoch": 3.2714285714285714, |
| "grad_norm": 0.4372345209121704, |
| "learning_rate": 3.483253588516747e-06, |
| "loss": 0.03564683347940445, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.03629, |
| "step": 687, |
| "tokens/total": 2813952, |
| "tokens/train_per_sec_per_gpu": 16.3, |
| "tokens/trainable": 201346 |
| }, |
| { |
| "epoch": 3.276190476190476, |
| "grad_norm": 0.5511652231216431, |
| "learning_rate": 3.473684210526316e-06, |
| "loss": 0.06980352848768234, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0723, |
| "step": 688, |
| "tokens/total": 2818048, |
| "tokens/train_per_sec_per_gpu": 14.69, |
| "tokens/trainable": 201643 |
| }, |
| { |
| "epoch": 3.280952380952381, |
| "grad_norm": 0.4855141043663025, |
| "learning_rate": 3.464114832535885e-06, |
| "loss": 0.04140230640769005, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04227, |
| "step": 689, |
| "tokens/total": 2822144, |
| "tokens/train_per_sec_per_gpu": 12.93, |
| "tokens/trainable": 201928 |
| }, |
| { |
| "epoch": 3.2857142857142856, |
| "grad_norm": 0.625236451625824, |
| "learning_rate": 3.454545454545455e-06, |
| "loss": 0.03237142413854599, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0329, |
| "step": 690, |
| "tokens/total": 2826240, |
| "tokens/train_per_sec_per_gpu": 12.73, |
| "tokens/trainable": 202176 |
| }, |
| { |
| "epoch": 3.2904761904761903, |
| "grad_norm": 0.5515985488891602, |
| "learning_rate": 3.4449760765550244e-06, |
| "loss": 0.0465005561709404, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0476, |
| "step": 691, |
| "tokens/total": 2830336, |
| "tokens/train_per_sec_per_gpu": 17.29, |
| "tokens/trainable": 202480 |
| }, |
| { |
| "epoch": 3.295238095238095, |
| "grad_norm": 0.48744669556617737, |
| "learning_rate": 3.4354066985645933e-06, |
| "loss": 0.04794064164161682, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.04911, |
| "step": 692, |
| "tokens/total": 2834432, |
| "tokens/train_per_sec_per_gpu": 13.07, |
| "tokens/trainable": 202752 |
| }, |
| { |
| "epoch": 3.3, |
| "grad_norm": 0.7841635346412659, |
| "learning_rate": 3.425837320574163e-06, |
| "loss": 0.10598424822092056, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1118, |
| "step": 693, |
| "tokens/total": 2838528, |
| "tokens/train_per_sec_per_gpu": 24.09, |
| "tokens/trainable": 203158 |
| }, |
| { |
| "epoch": 3.3047619047619046, |
| "grad_norm": 0.8256274461746216, |
| "learning_rate": 3.4162679425837325e-06, |
| "loss": 0.07799224555492401, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08111, |
| "step": 694, |
| "tokens/total": 2842624, |
| "tokens/train_per_sec_per_gpu": 9.86, |
| "tokens/trainable": 203449 |
| }, |
| { |
| "epoch": 3.3095238095238093, |
| "grad_norm": 0.9331197738647461, |
| "learning_rate": 3.4066985645933015e-06, |
| "loss": 0.07394289970397949, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07675, |
| "step": 695, |
| "tokens/total": 2846720, |
| "tokens/train_per_sec_per_gpu": 10.05, |
| "tokens/trainable": 203706 |
| }, |
| { |
| "epoch": 3.314285714285714, |
| "grad_norm": 0.4485538601875305, |
| "learning_rate": 3.397129186602871e-06, |
| "loss": 0.04761147126555443, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04876, |
| "step": 696, |
| "tokens/total": 2850816, |
| "tokens/train_per_sec_per_gpu": 17.09, |
| "tokens/trainable": 204025 |
| }, |
| { |
| "epoch": 3.319047619047619, |
| "grad_norm": 1.1187976598739624, |
| "learning_rate": 3.3875598086124407e-06, |
| "loss": 0.05788331478834152, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05959, |
| "step": 697, |
| "tokens/total": 2854912, |
| "tokens/train_per_sec_per_gpu": 10.64, |
| "tokens/trainable": 204293 |
| }, |
| { |
| "epoch": 3.323809523809524, |
| "grad_norm": 0.3998170495033264, |
| "learning_rate": 3.3779904306220097e-06, |
| "loss": 0.026133496314287186, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.02648, |
| "step": 698, |
| "tokens/total": 2859008, |
| "tokens/train_per_sec_per_gpu": 10.62, |
| "tokens/trainable": 204557 |
| }, |
| { |
| "epoch": 3.3285714285714287, |
| "grad_norm": 0.9097269773483276, |
| "learning_rate": 3.368421052631579e-06, |
| "loss": 0.12403593957424164, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.13206, |
| "step": 699, |
| "tokens/total": 2863104, |
| "tokens/train_per_sec_per_gpu": 12.79, |
| "tokens/trainable": 204830 |
| }, |
| { |
| "epoch": 3.3333333333333335, |
| "grad_norm": 0.5364436507225037, |
| "learning_rate": 3.358851674641149e-06, |
| "loss": 0.04892924055457115, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05015, |
| "step": 700, |
| "tokens/total": 2867200, |
| "tokens/train_per_sec_per_gpu": 11.38, |
| "tokens/trainable": 205084 |
| }, |
| { |
| "epoch": 3.3380952380952382, |
| "grad_norm": 1.4206129312515259, |
| "learning_rate": 3.3492822966507182e-06, |
| "loss": 0.06274130195379257, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06475, |
| "step": 701, |
| "tokens/total": 2871296, |
| "tokens/train_per_sec_per_gpu": 10.72, |
| "tokens/trainable": 205377 |
| }, |
| { |
| "epoch": 3.342857142857143, |
| "grad_norm": 0.7458651065826416, |
| "learning_rate": 3.339712918660287e-06, |
| "loss": 0.04512277990579605, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04616, |
| "step": 702, |
| "tokens/total": 2875392, |
| "tokens/train_per_sec_per_gpu": 11.81, |
| "tokens/trainable": 205627 |
| }, |
| { |
| "epoch": 3.3476190476190477, |
| "grad_norm": 0.8281404972076416, |
| "learning_rate": 3.3301435406698566e-06, |
| "loss": 0.07591881603002548, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07887, |
| "step": 703, |
| "tokens/total": 2879488, |
| "tokens/train_per_sec_per_gpu": 12.93, |
| "tokens/trainable": 205873 |
| }, |
| { |
| "epoch": 3.3523809523809525, |
| "grad_norm": 0.9755316972732544, |
| "learning_rate": 3.3205741626794264e-06, |
| "loss": 0.06004408001899719, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06188, |
| "step": 704, |
| "tokens/total": 2883584, |
| "tokens/train_per_sec_per_gpu": 15.18, |
| "tokens/trainable": 206183 |
| }, |
| { |
| "epoch": 3.357142857142857, |
| "grad_norm": 0.7920077443122864, |
| "learning_rate": 3.3110047846889954e-06, |
| "loss": 0.08305348455905914, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0866, |
| "step": 705, |
| "tokens/total": 2887680, |
| "tokens/train_per_sec_per_gpu": 13.99, |
| "tokens/trainable": 206461 |
| }, |
| { |
| "epoch": 3.361904761904762, |
| "grad_norm": 0.8308439254760742, |
| "learning_rate": 3.3014354066985648e-06, |
| "loss": 0.09610787779092789, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10088, |
| "step": 706, |
| "tokens/total": 2891776, |
| "tokens/train_per_sec_per_gpu": 14.53, |
| "tokens/trainable": 206834 |
| }, |
| { |
| "epoch": 3.3666666666666667, |
| "grad_norm": 0.7181533575057983, |
| "learning_rate": 3.2918660287081346e-06, |
| "loss": 0.054804570972919464, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05633, |
| "step": 707, |
| "tokens/total": 2895872, |
| "tokens/train_per_sec_per_gpu": 12.42, |
| "tokens/trainable": 207054 |
| }, |
| { |
| "epoch": 3.3714285714285714, |
| "grad_norm": 0.5962000489234924, |
| "learning_rate": 3.2822966507177035e-06, |
| "loss": 0.05790537968277931, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05961, |
| "step": 708, |
| "tokens/total": 2899968, |
| "tokens/train_per_sec_per_gpu": 13.25, |
| "tokens/trainable": 207343 |
| }, |
| { |
| "epoch": 3.376190476190476, |
| "grad_norm": 0.684431791305542, |
| "learning_rate": 3.272727272727273e-06, |
| "loss": 0.037118710577487946, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.03782, |
| "step": 709, |
| "tokens/total": 2904064, |
| "tokens/train_per_sec_per_gpu": 10.92, |
| "tokens/trainable": 207636 |
| }, |
| { |
| "epoch": 3.380952380952381, |
| "grad_norm": 0.8779307007789612, |
| "learning_rate": 3.2631578947368423e-06, |
| "loss": 0.06922551244497299, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07168, |
| "step": 710, |
| "tokens/total": 2908160, |
| "tokens/train_per_sec_per_gpu": 14.77, |
| "tokens/trainable": 207938 |
| }, |
| { |
| "epoch": 3.3857142857142857, |
| "grad_norm": 0.8065972328186035, |
| "learning_rate": 3.2535885167464117e-06, |
| "loss": 0.09936165809631348, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10447, |
| "step": 711, |
| "tokens/total": 2912256, |
| "tokens/train_per_sec_per_gpu": 25.22, |
| "tokens/trainable": 208414 |
| }, |
| { |
| "epoch": 3.3904761904761904, |
| "grad_norm": 0.6644219756126404, |
| "learning_rate": 3.244019138755981e-06, |
| "loss": 0.05635654181241989, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05797, |
| "step": 712, |
| "tokens/total": 2916352, |
| "tokens/train_per_sec_per_gpu": 15.97, |
| "tokens/trainable": 208705 |
| }, |
| { |
| "epoch": 3.395238095238095, |
| "grad_norm": 1.0275393724441528, |
| "learning_rate": 3.2344497607655505e-06, |
| "loss": 0.09682406485080719, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10167, |
| "step": 713, |
| "tokens/total": 2920448, |
| "tokens/train_per_sec_per_gpu": 17.09, |
| "tokens/trainable": 208987 |
| }, |
| { |
| "epoch": 3.4, |
| "grad_norm": 0.5890900492668152, |
| "learning_rate": 3.2248803827751194e-06, |
| "loss": 0.05707221478223801, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05873, |
| "step": 714, |
| "tokens/total": 2924544, |
| "tokens/train_per_sec_per_gpu": 15.93, |
| "tokens/trainable": 209293 |
| }, |
| { |
| "epoch": 3.4047619047619047, |
| "grad_norm": 0.9608796238899231, |
| "learning_rate": 3.2153110047846893e-06, |
| "loss": 0.07874710857868195, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08193, |
| "step": 715, |
| "tokens/total": 2928640, |
| "tokens/train_per_sec_per_gpu": 9.25, |
| "tokens/trainable": 209521 |
| }, |
| { |
| "epoch": 3.4095238095238094, |
| "grad_norm": 0.746338427066803, |
| "learning_rate": 3.2057416267942586e-06, |
| "loss": 0.09480582922697067, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09945, |
| "step": 716, |
| "tokens/total": 2932736, |
| "tokens/train_per_sec_per_gpu": 20.98, |
| "tokens/trainable": 209830 |
| }, |
| { |
| "epoch": 3.414285714285714, |
| "grad_norm": 0.7151622772216797, |
| "learning_rate": 3.1961722488038285e-06, |
| "loss": 0.07184720039367676, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07449, |
| "step": 717, |
| "tokens/total": 2936832, |
| "tokens/train_per_sec_per_gpu": 11.52, |
| "tokens/trainable": 210067 |
| }, |
| { |
| "epoch": 3.419047619047619, |
| "grad_norm": 1.9423059225082397, |
| "learning_rate": 3.1866028708133974e-06, |
| "loss": 0.08544290065765381, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0892, |
| "step": 718, |
| "tokens/total": 2940928, |
| "tokens/train_per_sec_per_gpu": 14.37, |
| "tokens/trainable": 210392 |
| }, |
| { |
| "epoch": 3.4238095238095236, |
| "grad_norm": 0.6730396747589111, |
| "learning_rate": 3.177033492822967e-06, |
| "loss": 0.08060389012098312, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08394, |
| "step": 719, |
| "tokens/total": 2945024, |
| "tokens/train_per_sec_per_gpu": 13.33, |
| "tokens/trainable": 210699 |
| }, |
| { |
| "epoch": 3.4285714285714284, |
| "grad_norm": 0.7929803729057312, |
| "learning_rate": 3.167464114832536e-06, |
| "loss": 0.11171211302280426, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11819, |
| "step": 720, |
| "tokens/total": 2949120, |
| "tokens/train_per_sec_per_gpu": 12.58, |
| "tokens/trainable": 210987 |
| }, |
| { |
| "epoch": 3.4333333333333336, |
| "grad_norm": 0.6450818777084351, |
| "learning_rate": 3.157894736842105e-06, |
| "loss": 0.07563444972038269, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07857, |
| "step": 721, |
| "tokens/total": 2953216, |
| "tokens/train_per_sec_per_gpu": 15.47, |
| "tokens/trainable": 211321 |
| }, |
| { |
| "epoch": 3.4380952380952383, |
| "grad_norm": 0.6084672212600708, |
| "learning_rate": 3.148325358851675e-06, |
| "loss": 0.03685470297932625, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03754, |
| "step": 722, |
| "tokens/total": 2957312, |
| "tokens/train_per_sec_per_gpu": 10.53, |
| "tokens/trainable": 211543 |
| }, |
| { |
| "epoch": 3.442857142857143, |
| "grad_norm": 1.0092453956604004, |
| "learning_rate": 3.1387559808612444e-06, |
| "loss": 0.07521015405654907, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07811, |
| "step": 723, |
| "tokens/total": 2961408, |
| "tokens/train_per_sec_per_gpu": 11.51, |
| "tokens/trainable": 211875 |
| }, |
| { |
| "epoch": 3.447619047619048, |
| "grad_norm": 0.5857625603675842, |
| "learning_rate": 3.1291866028708133e-06, |
| "loss": 0.06451966613531113, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06665, |
| "step": 724, |
| "tokens/total": 2965504, |
| "tokens/train_per_sec_per_gpu": 15.51, |
| "tokens/trainable": 212170 |
| }, |
| { |
| "epoch": 3.4523809523809526, |
| "grad_norm": 1.5166327953338623, |
| "learning_rate": 3.119617224880383e-06, |
| "loss": 0.04059155285358429, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.04143, |
| "step": 725, |
| "tokens/total": 2969600, |
| "tokens/train_per_sec_per_gpu": 14.51, |
| "tokens/trainable": 212427 |
| }, |
| { |
| "epoch": 3.4571428571428573, |
| "grad_norm": 0.5359680652618408, |
| "learning_rate": 3.1100478468899525e-06, |
| "loss": 0.05074106901884079, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05205, |
| "step": 726, |
| "tokens/total": 2973696, |
| "tokens/train_per_sec_per_gpu": 13.29, |
| "tokens/trainable": 212712 |
| }, |
| { |
| "epoch": 3.461904761904762, |
| "grad_norm": 0.5468071699142456, |
| "learning_rate": 3.1004784688995215e-06, |
| "loss": 0.04880739748477936, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05002, |
| "step": 727, |
| "tokens/total": 2977792, |
| "tokens/train_per_sec_per_gpu": 14.33, |
| "tokens/trainable": 212974 |
| }, |
| { |
| "epoch": 3.466666666666667, |
| "grad_norm": 0.743224024772644, |
| "learning_rate": 3.090909090909091e-06, |
| "loss": 0.07080194354057312, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07337, |
| "step": 728, |
| "tokens/total": 2981888, |
| "tokens/train_per_sec_per_gpu": 12.39, |
| "tokens/trainable": 213256 |
| }, |
| { |
| "epoch": 3.4714285714285715, |
| "grad_norm": 0.7840217351913452, |
| "learning_rate": 3.0813397129186607e-06, |
| "loss": 0.032732173800468445, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03327, |
| "step": 729, |
| "tokens/total": 2985984, |
| "tokens/train_per_sec_per_gpu": 11.76, |
| "tokens/trainable": 213529 |
| }, |
| { |
| "epoch": 3.4761904761904763, |
| "grad_norm": 0.7366862297058105, |
| "learning_rate": 3.07177033492823e-06, |
| "loss": 0.07588684558868408, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07884, |
| "step": 730, |
| "tokens/total": 2990080, |
| "tokens/train_per_sec_per_gpu": 12.81, |
| "tokens/trainable": 213807 |
| }, |
| { |
| "epoch": 3.480952380952381, |
| "grad_norm": 1.0238080024719238, |
| "learning_rate": 3.062200956937799e-06, |
| "loss": 0.09773441404104233, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10267, |
| "step": 731, |
| "tokens/total": 2994176, |
| "tokens/train_per_sec_per_gpu": 12.9, |
| "tokens/trainable": 214101 |
| }, |
| { |
| "epoch": 3.4857142857142858, |
| "grad_norm": 0.5155897736549377, |
| "learning_rate": 3.052631578947369e-06, |
| "loss": 0.03965584188699722, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04045, |
| "step": 732, |
| "tokens/total": 2998272, |
| "tokens/train_per_sec_per_gpu": 13.28, |
| "tokens/trainable": 214399 |
| }, |
| { |
| "epoch": 3.4904761904761905, |
| "grad_norm": 0.41440004110336304, |
| "learning_rate": 3.0430622009569383e-06, |
| "loss": 0.04801607131958008, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04919, |
| "step": 733, |
| "tokens/total": 3002368, |
| "tokens/train_per_sec_per_gpu": 18.11, |
| "tokens/trainable": 214776 |
| }, |
| { |
| "epoch": 3.4952380952380953, |
| "grad_norm": 0.7627658247947693, |
| "learning_rate": 3.0334928229665072e-06, |
| "loss": 0.09889574348926544, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.10395, |
| "step": 734, |
| "tokens/total": 3006464, |
| "tokens/train_per_sec_per_gpu": 18.78, |
| "tokens/trainable": 215155 |
| }, |
| { |
| "epoch": 3.5, |
| "grad_norm": 0.7029794454574585, |
| "learning_rate": 3.0239234449760766e-06, |
| "loss": 0.09765593707561493, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10258, |
| "step": 735, |
| "tokens/total": 3010560, |
| "tokens/train_per_sec_per_gpu": 20.18, |
| "tokens/trainable": 215539 |
| }, |
| { |
| "epoch": 3.5047619047619047, |
| "grad_norm": 1.4511120319366455, |
| "learning_rate": 3.0143540669856464e-06, |
| "loss": 0.1049220860004425, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11062, |
| "step": 736, |
| "tokens/total": 3014656, |
| "tokens/train_per_sec_per_gpu": 10.55, |
| "tokens/trainable": 215790 |
| }, |
| { |
| "epoch": 3.5095238095238095, |
| "grad_norm": 0.6316392421722412, |
| "learning_rate": 3.0047846889952154e-06, |
| "loss": 0.05280493572354317, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05422, |
| "step": 737, |
| "tokens/total": 3018752, |
| "tokens/train_per_sec_per_gpu": 11.14, |
| "tokens/trainable": 216068 |
| }, |
| { |
| "epoch": 3.5142857142857142, |
| "grad_norm": 0.8526849150657654, |
| "learning_rate": 2.9952153110047848e-06, |
| "loss": 0.09136175364255905, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.09567, |
| "step": 738, |
| "tokens/total": 3022848, |
| "tokens/train_per_sec_per_gpu": 19.51, |
| "tokens/trainable": 216392 |
| }, |
| { |
| "epoch": 3.519047619047619, |
| "grad_norm": 0.6972112655639648, |
| "learning_rate": 2.9856459330143546e-06, |
| "loss": 0.06299304217100143, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06502, |
| "step": 739, |
| "tokens/total": 3026944, |
| "tokens/train_per_sec_per_gpu": 7.72, |
| "tokens/trainable": 216620 |
| }, |
| { |
| "epoch": 3.5238095238095237, |
| "grad_norm": 1.0006465911865234, |
| "learning_rate": 2.9760765550239235e-06, |
| "loss": 0.1156340166926384, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12258, |
| "step": 740, |
| "tokens/total": 3031040, |
| "tokens/train_per_sec_per_gpu": 14.99, |
| "tokens/trainable": 216946 |
| }, |
| { |
| "epoch": 3.5285714285714285, |
| "grad_norm": 1.0956883430480957, |
| "learning_rate": 2.966507177033493e-06, |
| "loss": 0.07261288166046143, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07531, |
| "step": 741, |
| "tokens/total": 3035136, |
| "tokens/train_per_sec_per_gpu": 11.77, |
| "tokens/trainable": 217231 |
| }, |
| { |
| "epoch": 3.533333333333333, |
| "grad_norm": 1.1765693426132202, |
| "learning_rate": 2.9569377990430627e-06, |
| "loss": 0.08260992914438248, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08612, |
| "step": 742, |
| "tokens/total": 3039232, |
| "tokens/train_per_sec_per_gpu": 10.25, |
| "tokens/trainable": 217490 |
| }, |
| { |
| "epoch": 3.538095238095238, |
| "grad_norm": 0.9932781457901001, |
| "learning_rate": 2.9473684210526317e-06, |
| "loss": 0.11413135379552841, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.1209, |
| "step": 743, |
| "tokens/total": 3043328, |
| "tokens/train_per_sec_per_gpu": 12.99, |
| "tokens/trainable": 217787 |
| }, |
| { |
| "epoch": 3.5428571428571427, |
| "grad_norm": 0.34685567021369934, |
| "learning_rate": 2.937799043062201e-06, |
| "loss": 0.028304440900683403, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02871, |
| "step": 744, |
| "tokens/total": 3047424, |
| "tokens/train_per_sec_per_gpu": 16.11, |
| "tokens/trainable": 218083 |
| }, |
| { |
| "epoch": 3.5476190476190474, |
| "grad_norm": 0.8639155626296997, |
| "learning_rate": 2.9282296650717705e-06, |
| "loss": 0.06153899058699608, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06347, |
| "step": 745, |
| "tokens/total": 3051520, |
| "tokens/train_per_sec_per_gpu": 13.48, |
| "tokens/trainable": 218376 |
| }, |
| { |
| "epoch": 3.552380952380952, |
| "grad_norm": 0.7670663595199585, |
| "learning_rate": 2.9186602870813403e-06, |
| "loss": 0.05310921370983124, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05454, |
| "step": 746, |
| "tokens/total": 3055616, |
| "tokens/train_per_sec_per_gpu": 9.47, |
| "tokens/trainable": 218617 |
| }, |
| { |
| "epoch": 3.557142857142857, |
| "grad_norm": 0.7092881202697754, |
| "learning_rate": 2.9090909090909093e-06, |
| "loss": 0.05616293102502823, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05777, |
| "step": 747, |
| "tokens/total": 3059712, |
| "tokens/train_per_sec_per_gpu": 19.82, |
| "tokens/trainable": 218932 |
| }, |
| { |
| "epoch": 3.5619047619047617, |
| "grad_norm": 0.6593560576438904, |
| "learning_rate": 2.8995215311004787e-06, |
| "loss": 0.06515651196241379, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06733, |
| "step": 748, |
| "tokens/total": 3063808, |
| "tokens/train_per_sec_per_gpu": 15.37, |
| "tokens/trainable": 219262 |
| }, |
| { |
| "epoch": 3.5666666666666664, |
| "grad_norm": 0.9004426598548889, |
| "learning_rate": 2.8899521531100485e-06, |
| "loss": 0.08489474654197693, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0886, |
| "step": 749, |
| "tokens/total": 3067904, |
| "tokens/train_per_sec_per_gpu": 11.18, |
| "tokens/trainable": 219472 |
| }, |
| { |
| "epoch": 3.571428571428571, |
| "grad_norm": 0.9647462964057922, |
| "learning_rate": 2.8803827751196174e-06, |
| "loss": 0.08201523125171661, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08547, |
| "step": 750, |
| "tokens/total": 3072000, |
| "tokens/train_per_sec_per_gpu": 12.37, |
| "tokens/trainable": 219738 |
| }, |
| { |
| "epoch": 3.576190476190476, |
| "grad_norm": 0.4528436064720154, |
| "learning_rate": 2.870813397129187e-06, |
| "loss": 0.03055403381586075, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03103, |
| "step": 751, |
| "tokens/total": 3076096, |
| "tokens/train_per_sec_per_gpu": 9.11, |
| "tokens/trainable": 220017 |
| }, |
| { |
| "epoch": 3.580952380952381, |
| "grad_norm": 0.7963530421257019, |
| "learning_rate": 2.861244019138756e-06, |
| "loss": 0.08453945815563202, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08822, |
| "step": 752, |
| "tokens/total": 3080192, |
| "tokens/train_per_sec_per_gpu": 20.97, |
| "tokens/trainable": 220419 |
| }, |
| { |
| "epoch": 3.585714285714286, |
| "grad_norm": 0.8981921076774597, |
| "learning_rate": 2.851674641148325e-06, |
| "loss": 0.09118165075778961, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09547, |
| "step": 753, |
| "tokens/total": 3084288, |
| "tokens/train_per_sec_per_gpu": 11.72, |
| "tokens/trainable": 220827 |
| }, |
| { |
| "epoch": 3.5904761904761906, |
| "grad_norm": 0.5238147974014282, |
| "learning_rate": 2.842105263157895e-06, |
| "loss": 0.04031630605459213, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04114, |
| "step": 754, |
| "tokens/total": 3088384, |
| "tokens/train_per_sec_per_gpu": 16.61, |
| "tokens/trainable": 221140 |
| }, |
| { |
| "epoch": 3.5952380952380953, |
| "grad_norm": 0.9348918795585632, |
| "learning_rate": 2.8325358851674644e-06, |
| "loss": 0.0727066695690155, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07542, |
| "step": 755, |
| "tokens/total": 3092480, |
| "tokens/train_per_sec_per_gpu": 17.87, |
| "tokens/trainable": 221519 |
| }, |
| { |
| "epoch": 3.6, |
| "grad_norm": 0.7450806498527527, |
| "learning_rate": 2.8229665071770333e-06, |
| "loss": 0.07027944177389145, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07281, |
| "step": 756, |
| "tokens/total": 3096576, |
| "tokens/train_per_sec_per_gpu": 12.8, |
| "tokens/trainable": 221815 |
| }, |
| { |
| "epoch": 3.604761904761905, |
| "grad_norm": 0.8082990050315857, |
| "learning_rate": 2.813397129186603e-06, |
| "loss": 0.07638609409332275, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07938, |
| "step": 757, |
| "tokens/total": 3100672, |
| "tokens/train_per_sec_per_gpu": 13.99, |
| "tokens/trainable": 222138 |
| }, |
| { |
| "epoch": 3.6095238095238096, |
| "grad_norm": 1.0129432678222656, |
| "learning_rate": 2.8038277511961725e-06, |
| "loss": 0.12862154841423035, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13726, |
| "step": 758, |
| "tokens/total": 3104768, |
| "tokens/train_per_sec_per_gpu": 11.76, |
| "tokens/trainable": 222473 |
| }, |
| { |
| "epoch": 3.6142857142857143, |
| "grad_norm": 0.42732250690460205, |
| "learning_rate": 2.794258373205742e-06, |
| "loss": 0.05144212022423744, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05279, |
| "step": 759, |
| "tokens/total": 3108864, |
| "tokens/train_per_sec_per_gpu": 15.42, |
| "tokens/trainable": 222800 |
| }, |
| { |
| "epoch": 3.619047619047619, |
| "grad_norm": 0.4792739450931549, |
| "learning_rate": 2.7846889952153113e-06, |
| "loss": 0.0455373115837574, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04659, |
| "step": 760, |
| "tokens/total": 3112960, |
| "tokens/train_per_sec_per_gpu": 8.68, |
| "tokens/trainable": 223081 |
| }, |
| { |
| "epoch": 3.623809523809524, |
| "grad_norm": 1.2429322004318237, |
| "learning_rate": 2.7751196172248807e-06, |
| "loss": 0.09872013330459595, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10376, |
| "step": 761, |
| "tokens/total": 3117056, |
| "tokens/train_per_sec_per_gpu": 12.91, |
| "tokens/trainable": 223343 |
| }, |
| { |
| "epoch": 3.6285714285714286, |
| "grad_norm": 1.2431164979934692, |
| "learning_rate": 2.76555023923445e-06, |
| "loss": 0.09223204851150513, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09662, |
| "step": 762, |
| "tokens/total": 3121152, |
| "tokens/train_per_sec_per_gpu": 9.73, |
| "tokens/trainable": 223573 |
| }, |
| { |
| "epoch": 3.6333333333333333, |
| "grad_norm": 0.5852113962173462, |
| "learning_rate": 2.755980861244019e-06, |
| "loss": 0.09133613854646683, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.09564, |
| "step": 763, |
| "tokens/total": 3125248, |
| "tokens/train_per_sec_per_gpu": 15.86, |
| "tokens/trainable": 223932 |
| }, |
| { |
| "epoch": 3.638095238095238, |
| "grad_norm": 0.7849932312965393, |
| "learning_rate": 2.746411483253589e-06, |
| "loss": 0.08221340924501419, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08569, |
| "step": 764, |
| "tokens/total": 3129344, |
| "tokens/train_per_sec_per_gpu": 12.63, |
| "tokens/trainable": 224255 |
| }, |
| { |
| "epoch": 3.642857142857143, |
| "grad_norm": 0.8393433690071106, |
| "learning_rate": 2.7368421052631583e-06, |
| "loss": 0.08088158071041107, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08424, |
| "step": 765, |
| "tokens/total": 3133440, |
| "tokens/train_per_sec_per_gpu": 14.01, |
| "tokens/trainable": 224537 |
| }, |
| { |
| "epoch": 3.6476190476190475, |
| "grad_norm": 1.1545741558074951, |
| "learning_rate": 2.7272727272727272e-06, |
| "loss": 0.11216983199119568, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1187, |
| "step": 766, |
| "tokens/total": 3137536, |
| "tokens/train_per_sec_per_gpu": 17.2, |
| "tokens/trainable": 224863 |
| }, |
| { |
| "epoch": 3.6523809523809523, |
| "grad_norm": 0.8742523193359375, |
| "learning_rate": 2.717703349282297e-06, |
| "loss": 0.10797283798456192, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11402, |
| "step": 767, |
| "tokens/total": 3141632, |
| "tokens/train_per_sec_per_gpu": 12.15, |
| "tokens/trainable": 225163 |
| }, |
| { |
| "epoch": 3.657142857142857, |
| "grad_norm": 0.7308986186981201, |
| "learning_rate": 2.7081339712918664e-06, |
| "loss": 0.110054612159729, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11634, |
| "step": 768, |
| "tokens/total": 3145728, |
| "tokens/train_per_sec_per_gpu": 15.87, |
| "tokens/trainable": 225510 |
| }, |
| { |
| "epoch": 3.6619047619047618, |
| "grad_norm": 0.7405397295951843, |
| "learning_rate": 2.6985645933014354e-06, |
| "loss": 0.10397686064243317, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10957, |
| "step": 769, |
| "tokens/total": 3149824, |
| "tokens/train_per_sec_per_gpu": 11.43, |
| "tokens/trainable": 225742 |
| }, |
| { |
| "epoch": 3.6666666666666665, |
| "grad_norm": 1.0262563228607178, |
| "learning_rate": 2.6889952153110048e-06, |
| "loss": 0.13542257249355316, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.14502, |
| "step": 770, |
| "tokens/total": 3153920, |
| "tokens/train_per_sec_per_gpu": 16.94, |
| "tokens/trainable": 226075 |
| }, |
| { |
| "epoch": 3.6714285714285713, |
| "grad_norm": 1.6212959289550781, |
| "learning_rate": 2.6794258373205746e-06, |
| "loss": 0.11226394772529602, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11881, |
| "step": 771, |
| "tokens/total": 3158016, |
| "tokens/train_per_sec_per_gpu": 11.9, |
| "tokens/trainable": 226296 |
| }, |
| { |
| "epoch": 3.6761904761904765, |
| "grad_norm": 1.3431205749511719, |
| "learning_rate": 2.669856459330144e-06, |
| "loss": 0.07553976774215698, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07847, |
| "step": 772, |
| "tokens/total": 3162112, |
| "tokens/train_per_sec_per_gpu": 10.88, |
| "tokens/trainable": 226564 |
| }, |
| { |
| "epoch": 3.680952380952381, |
| "grad_norm": 0.9042708873748779, |
| "learning_rate": 2.660287081339713e-06, |
| "loss": 0.061304397881031036, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06322, |
| "step": 773, |
| "tokens/total": 3166208, |
| "tokens/train_per_sec_per_gpu": 11.21, |
| "tokens/trainable": 226906 |
| }, |
| { |
| "epoch": 3.685714285714286, |
| "grad_norm": 0.7307568192481995, |
| "learning_rate": 2.6507177033492828e-06, |
| "loss": 0.0648641288280487, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06701, |
| "step": 774, |
| "tokens/total": 3170304, |
| "tokens/train_per_sec_per_gpu": 12.27, |
| "tokens/trainable": 227198 |
| }, |
| { |
| "epoch": 3.6904761904761907, |
| "grad_norm": 0.8105649352073669, |
| "learning_rate": 2.641148325358852e-06, |
| "loss": 0.06054212898015976, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06241, |
| "step": 775, |
| "tokens/total": 3174400, |
| "tokens/train_per_sec_per_gpu": 9.69, |
| "tokens/trainable": 227476 |
| }, |
| { |
| "epoch": 3.6952380952380954, |
| "grad_norm": 1.0654374361038208, |
| "learning_rate": 2.631578947368421e-06, |
| "loss": 0.06407278776168823, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06617, |
| "step": 776, |
| "tokens/total": 3178496, |
| "tokens/train_per_sec_per_gpu": 9.43, |
| "tokens/trainable": 227697 |
| }, |
| { |
| "epoch": 3.7, |
| "grad_norm": 1.0169827938079834, |
| "learning_rate": 2.6220095693779905e-06, |
| "loss": 0.12414013594388962, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13217, |
| "step": 777, |
| "tokens/total": 3182592, |
| "tokens/train_per_sec_per_gpu": 8.18, |
| "tokens/trainable": 227923 |
| }, |
| { |
| "epoch": 3.704761904761905, |
| "grad_norm": 0.8178273439407349, |
| "learning_rate": 2.6124401913875603e-06, |
| "loss": 0.08592554926872253, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08973, |
| "step": 778, |
| "tokens/total": 3186688, |
| "tokens/train_per_sec_per_gpu": 14.14, |
| "tokens/trainable": 228285 |
| }, |
| { |
| "epoch": 3.7095238095238097, |
| "grad_norm": 0.6456396579742432, |
| "learning_rate": 2.6028708133971293e-06, |
| "loss": 0.05404656380414963, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05553, |
| "step": 779, |
| "tokens/total": 3190784, |
| "tokens/train_per_sec_per_gpu": 11.12, |
| "tokens/trainable": 228537 |
| }, |
| { |
| "epoch": 3.7142857142857144, |
| "grad_norm": 1.2505536079406738, |
| "learning_rate": 2.5933014354066987e-06, |
| "loss": 0.08313541859388351, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08669, |
| "step": 780, |
| "tokens/total": 3194880, |
| "tokens/train_per_sec_per_gpu": 11.8, |
| "tokens/trainable": 228765 |
| }, |
| { |
| "epoch": 3.719047619047619, |
| "grad_norm": 0.4605754613876343, |
| "learning_rate": 2.5837320574162685e-06, |
| "loss": 0.05742780119180679, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05911, |
| "step": 781, |
| "tokens/total": 3198976, |
| "tokens/train_per_sec_per_gpu": 16.55, |
| "tokens/trainable": 229136 |
| }, |
| { |
| "epoch": 3.723809523809524, |
| "grad_norm": 0.7865471243858337, |
| "learning_rate": 2.5741626794258374e-06, |
| "loss": 0.09312713891267776, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0976, |
| "step": 782, |
| "tokens/total": 3203072, |
| "tokens/train_per_sec_per_gpu": 12.28, |
| "tokens/trainable": 229397 |
| }, |
| { |
| "epoch": 3.7285714285714286, |
| "grad_norm": 0.4759027659893036, |
| "learning_rate": 2.564593301435407e-06, |
| "loss": 0.022237185388803482, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02249, |
| "step": 783, |
| "tokens/total": 3207168, |
| "tokens/train_per_sec_per_gpu": 11.76, |
| "tokens/trainable": 229670 |
| }, |
| { |
| "epoch": 3.7333333333333334, |
| "grad_norm": 0.6794772148132324, |
| "learning_rate": 2.5550239234449762e-06, |
| "loss": 0.05335363745689392, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0548, |
| "step": 784, |
| "tokens/total": 3211264, |
| "tokens/train_per_sec_per_gpu": 11.65, |
| "tokens/trainable": 229934 |
| }, |
| { |
| "epoch": 3.738095238095238, |
| "grad_norm": 0.859467089176178, |
| "learning_rate": 2.5454545454545456e-06, |
| "loss": 0.0922853946685791, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09668, |
| "step": 785, |
| "tokens/total": 3215360, |
| "tokens/train_per_sec_per_gpu": 11.79, |
| "tokens/trainable": 230189 |
| }, |
| { |
| "epoch": 3.742857142857143, |
| "grad_norm": 1.0541110038757324, |
| "learning_rate": 2.535885167464115e-06, |
| "loss": 0.1113077700138092, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11774, |
| "step": 786, |
| "tokens/total": 3219456, |
| "tokens/train_per_sec_per_gpu": 18.41, |
| "tokens/trainable": 230510 |
| }, |
| { |
| "epoch": 3.7476190476190476, |
| "grad_norm": 0.8308860659599304, |
| "learning_rate": 2.5263157894736844e-06, |
| "loss": 0.07469616085290909, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07756, |
| "step": 787, |
| "tokens/total": 3223552, |
| "tokens/train_per_sec_per_gpu": 13.46, |
| "tokens/trainable": 230769 |
| }, |
| { |
| "epoch": 3.7523809523809524, |
| "grad_norm": 0.5026382207870483, |
| "learning_rate": 2.516746411483254e-06, |
| "loss": 0.04333016648888588, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04428, |
| "step": 788, |
| "tokens/total": 3227648, |
| "tokens/train_per_sec_per_gpu": 11.95, |
| "tokens/trainable": 231027 |
| }, |
| { |
| "epoch": 3.757142857142857, |
| "grad_norm": 0.8133578300476074, |
| "learning_rate": 2.507177033492823e-06, |
| "loss": 0.0831025242805481, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08665, |
| "step": 789, |
| "tokens/total": 3231744, |
| "tokens/train_per_sec_per_gpu": 17.59, |
| "tokens/trainable": 231302 |
| }, |
| { |
| "epoch": 3.761904761904762, |
| "grad_norm": 0.8333939909934998, |
| "learning_rate": 2.4976076555023926e-06, |
| "loss": 0.08183260262012482, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08527, |
| "step": 790, |
| "tokens/total": 3235840, |
| "tokens/train_per_sec_per_gpu": 13.2, |
| "tokens/trainable": 231553 |
| }, |
| { |
| "epoch": 3.7666666666666666, |
| "grad_norm": 0.845085084438324, |
| "learning_rate": 2.488038277511962e-06, |
| "loss": 0.07390055060386658, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.0767, |
| "step": 791, |
| "tokens/total": 3239936, |
| "tokens/train_per_sec_per_gpu": 24.47, |
| "tokens/trainable": 231977 |
| }, |
| { |
| "epoch": 3.7714285714285714, |
| "grad_norm": 0.7061827182769775, |
| "learning_rate": 2.4784688995215313e-06, |
| "loss": 0.07624020427465439, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07922, |
| "step": 792, |
| "tokens/total": 3244032, |
| "tokens/train_per_sec_per_gpu": 21.24, |
| "tokens/trainable": 232345 |
| }, |
| { |
| "epoch": 3.776190476190476, |
| "grad_norm": 0.6797322034835815, |
| "learning_rate": 2.4688995215311007e-06, |
| "loss": 0.06764702498912811, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06999, |
| "step": 793, |
| "tokens/total": 3248128, |
| "tokens/train_per_sec_per_gpu": 16.98, |
| "tokens/trainable": 232686 |
| }, |
| { |
| "epoch": 3.780952380952381, |
| "grad_norm": 0.8728348612785339, |
| "learning_rate": 2.45933014354067e-06, |
| "loss": 0.09084605425596237, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0951, |
| "step": 794, |
| "tokens/total": 3252224, |
| "tokens/train_per_sec_per_gpu": 21.61, |
| "tokens/trainable": 233037 |
| }, |
| { |
| "epoch": 3.7857142857142856, |
| "grad_norm": 0.48481041193008423, |
| "learning_rate": 2.4497607655502395e-06, |
| "loss": 0.07377579808235168, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07657, |
| "step": 795, |
| "tokens/total": 3256320, |
| "tokens/train_per_sec_per_gpu": 17.95, |
| "tokens/trainable": 233376 |
| }, |
| { |
| "epoch": 3.7904761904761903, |
| "grad_norm": 0.6856587529182434, |
| "learning_rate": 2.440191387559809e-06, |
| "loss": 0.08114641904830933, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08453, |
| "step": 796, |
| "tokens/total": 3260416, |
| "tokens/train_per_sec_per_gpu": 10.61, |
| "tokens/trainable": 233714 |
| }, |
| { |
| "epoch": 3.795238095238095, |
| "grad_norm": 0.7820855379104614, |
| "learning_rate": 2.4306220095693783e-06, |
| "loss": 0.07170091569423676, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07433, |
| "step": 797, |
| "tokens/total": 3264512, |
| "tokens/train_per_sec_per_gpu": 19.66, |
| "tokens/trainable": 234046 |
| }, |
| { |
| "epoch": 3.8, |
| "grad_norm": 1.2638566493988037, |
| "learning_rate": 2.4210526315789477e-06, |
| "loss": 0.10723811388015747, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1132, |
| "step": 798, |
| "tokens/total": 3268608, |
| "tokens/train_per_sec_per_gpu": 9.52, |
| "tokens/trainable": 234329 |
| }, |
| { |
| "epoch": 3.8047619047619046, |
| "grad_norm": 0.7651883959770203, |
| "learning_rate": 2.411483253588517e-06, |
| "loss": 0.09171386063098907, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.09605, |
| "step": 799, |
| "tokens/total": 3272704, |
| "tokens/train_per_sec_per_gpu": 16.78, |
| "tokens/trainable": 234646 |
| }, |
| { |
| "epoch": 3.8095238095238093, |
| "grad_norm": 0.5760342478752136, |
| "learning_rate": 2.4019138755980864e-06, |
| "loss": 0.04247612506151199, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04339, |
| "step": 800, |
| "tokens/total": 3276800, |
| "tokens/train_per_sec_per_gpu": 10.31, |
| "tokens/trainable": 234874 |
| }, |
| { |
| "epoch": 3.814285714285714, |
| "grad_norm": 0.43003764748573303, |
| "learning_rate": 2.392344497607656e-06, |
| "loss": 0.06375513225793839, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06583, |
| "step": 801, |
| "tokens/total": 3280896, |
| "tokens/train_per_sec_per_gpu": 8.48, |
| "tokens/trainable": 235204 |
| }, |
| { |
| "epoch": 3.819047619047619, |
| "grad_norm": 1.405259609222412, |
| "learning_rate": 2.382775119617225e-06, |
| "loss": 0.14087527990341187, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.15128, |
| "step": 802, |
| "tokens/total": 3284992, |
| "tokens/train_per_sec_per_gpu": 12.94, |
| "tokens/trainable": 235427 |
| }, |
| { |
| "epoch": 3.8238095238095235, |
| "grad_norm": 1.2408273220062256, |
| "learning_rate": 2.3732057416267946e-06, |
| "loss": 0.11416477710008621, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12094, |
| "step": 803, |
| "tokens/total": 3289088, |
| "tokens/train_per_sec_per_gpu": 9.93, |
| "tokens/trainable": 235675 |
| }, |
| { |
| "epoch": 3.8285714285714287, |
| "grad_norm": 0.7246369123458862, |
| "learning_rate": 2.363636363636364e-06, |
| "loss": 0.06474482268095016, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06689, |
| "step": 804, |
| "tokens/total": 3293184, |
| "tokens/train_per_sec_per_gpu": 12.8, |
| "tokens/trainable": 235969 |
| }, |
| { |
| "epoch": 3.8333333333333335, |
| "grad_norm": 1.1100120544433594, |
| "learning_rate": 2.3540669856459334e-06, |
| "loss": 0.10005244612693787, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10523, |
| "step": 805, |
| "tokens/total": 3297280, |
| "tokens/train_per_sec_per_gpu": 14.92, |
| "tokens/trainable": 236255 |
| }, |
| { |
| "epoch": 3.8380952380952382, |
| "grad_norm": 0.6413360238075256, |
| "learning_rate": 2.3444976076555028e-06, |
| "loss": 0.05007662996649742, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05135, |
| "step": 806, |
| "tokens/total": 3301376, |
| "tokens/train_per_sec_per_gpu": 11.21, |
| "tokens/trainable": 236534 |
| }, |
| { |
| "epoch": 3.842857142857143, |
| "grad_norm": 0.5169480443000793, |
| "learning_rate": 2.3349282296650717e-06, |
| "loss": 0.03839801624417305, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03914, |
| "step": 807, |
| "tokens/total": 3305472, |
| "tokens/train_per_sec_per_gpu": 12.93, |
| "tokens/trainable": 236885 |
| }, |
| { |
| "epoch": 3.8476190476190477, |
| "grad_norm": 0.7787706255912781, |
| "learning_rate": 2.3253588516746415e-06, |
| "loss": 0.08824636042118073, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09226, |
| "step": 808, |
| "tokens/total": 3309568, |
| "tokens/train_per_sec_per_gpu": 16.77, |
| "tokens/trainable": 237173 |
| }, |
| { |
| "epoch": 3.8523809523809525, |
| "grad_norm": 0.7283034324645996, |
| "learning_rate": 2.3157894736842105e-06, |
| "loss": 0.06264171004295349, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06465, |
| "step": 809, |
| "tokens/total": 3313664, |
| "tokens/train_per_sec_per_gpu": 9.94, |
| "tokens/trainable": 237473 |
| }, |
| { |
| "epoch": 3.857142857142857, |
| "grad_norm": 0.6716837882995605, |
| "learning_rate": 2.30622009569378e-06, |
| "loss": 0.07053433358669281, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07308, |
| "step": 810, |
| "tokens/total": 3317760, |
| "tokens/train_per_sec_per_gpu": 14.9, |
| "tokens/trainable": 237757 |
| }, |
| { |
| "epoch": 3.861904761904762, |
| "grad_norm": 0.6464008092880249, |
| "learning_rate": 2.2966507177033497e-06, |
| "loss": 0.07775363326072693, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08086, |
| "step": 811, |
| "tokens/total": 3321856, |
| "tokens/train_per_sec_per_gpu": 11.28, |
| "tokens/trainable": 238049 |
| }, |
| { |
| "epoch": 3.8666666666666667, |
| "grad_norm": 0.7403518557548523, |
| "learning_rate": 2.2870813397129187e-06, |
| "loss": 0.09577998518943787, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10052, |
| "step": 812, |
| "tokens/total": 3325952, |
| "tokens/train_per_sec_per_gpu": 12.57, |
| "tokens/trainable": 238325 |
| }, |
| { |
| "epoch": 3.8714285714285714, |
| "grad_norm": 0.9457064270973206, |
| "learning_rate": 2.2775119617224885e-06, |
| "loss": 0.03777073323726654, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03849, |
| "step": 813, |
| "tokens/total": 3330048, |
| "tokens/train_per_sec_per_gpu": 11.52, |
| "tokens/trainable": 238556 |
| }, |
| { |
| "epoch": 3.876190476190476, |
| "grad_norm": 0.6895711421966553, |
| "learning_rate": 2.2679425837320575e-06, |
| "loss": 0.06421350687742233, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06632, |
| "step": 814, |
| "tokens/total": 3334144, |
| "tokens/train_per_sec_per_gpu": 9.66, |
| "tokens/trainable": 238825 |
| }, |
| { |
| "epoch": 3.880952380952381, |
| "grad_norm": 1.067850947380066, |
| "learning_rate": 2.258373205741627e-06, |
| "loss": 0.12325073778629303, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.13117, |
| "step": 815, |
| "tokens/total": 3338240, |
| "tokens/train_per_sec_per_gpu": 11.6, |
| "tokens/trainable": 239107 |
| }, |
| { |
| "epoch": 3.8857142857142857, |
| "grad_norm": 0.9182291030883789, |
| "learning_rate": 2.2488038277511962e-06, |
| "loss": 0.06907776743173599, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07152, |
| "step": 816, |
| "tokens/total": 3342336, |
| "tokens/train_per_sec_per_gpu": 17.84, |
| "tokens/trainable": 239421 |
| }, |
| { |
| "epoch": 3.8904761904761904, |
| "grad_norm": 0.6250137090682983, |
| "learning_rate": 2.2392344497607656e-06, |
| "loss": 0.05608314275741577, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.05769, |
| "step": 817, |
| "tokens/total": 3346432, |
| "tokens/train_per_sec_per_gpu": 15.91, |
| "tokens/trainable": 239707 |
| }, |
| { |
| "epoch": 3.895238095238095, |
| "grad_norm": 0.8182249069213867, |
| "learning_rate": 2.229665071770335e-06, |
| "loss": 0.057830363512039185, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.05954, |
| "step": 818, |
| "tokens/total": 3350528, |
| "tokens/train_per_sec_per_gpu": 11.06, |
| "tokens/trainable": 239938 |
| }, |
| { |
| "epoch": 3.9, |
| "grad_norm": 0.8500362634658813, |
| "learning_rate": 2.2200956937799044e-06, |
| "loss": 0.11211861670017242, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11865, |
| "step": 819, |
| "tokens/total": 3354624, |
| "tokens/train_per_sec_per_gpu": 9.0, |
| "tokens/trainable": 240191 |
| }, |
| { |
| "epoch": 3.9047619047619047, |
| "grad_norm": 0.6874716281890869, |
| "learning_rate": 2.2105263157894738e-06, |
| "loss": 0.062088239938020706, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06406, |
| "step": 820, |
| "tokens/total": 3358720, |
| "tokens/train_per_sec_per_gpu": 12.66, |
| "tokens/trainable": 240485 |
| }, |
| { |
| "epoch": 3.9095238095238094, |
| "grad_norm": 0.8891911506652832, |
| "learning_rate": 2.200956937799043e-06, |
| "loss": 0.08611418306827545, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08993, |
| "step": 821, |
| "tokens/total": 3362816, |
| "tokens/train_per_sec_per_gpu": 10.57, |
| "tokens/trainable": 240753 |
| }, |
| { |
| "epoch": 3.914285714285714, |
| "grad_norm": 1.2441385984420776, |
| "learning_rate": 2.1913875598086126e-06, |
| "loss": 0.0724477618932724, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07514, |
| "step": 822, |
| "tokens/total": 3366912, |
| "tokens/train_per_sec_per_gpu": 8.75, |
| "tokens/trainable": 240943 |
| }, |
| { |
| "epoch": 3.919047619047619, |
| "grad_norm": 0.7042216062545776, |
| "learning_rate": 2.181818181818182e-06, |
| "loss": 0.08102297782897949, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0844, |
| "step": 823, |
| "tokens/total": 3371008, |
| "tokens/train_per_sec_per_gpu": 11.55, |
| "tokens/trainable": 241235 |
| }, |
| { |
| "epoch": 3.923809523809524, |
| "grad_norm": 5.344852924346924, |
| "learning_rate": 2.1722488038277513e-06, |
| "loss": 0.07976698130369186, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.08303, |
| "step": 824, |
| "tokens/total": 3375104, |
| "tokens/train_per_sec_per_gpu": 16.83, |
| "tokens/trainable": 241603 |
| }, |
| { |
| "epoch": 3.928571428571429, |
| "grad_norm": 0.492816299200058, |
| "learning_rate": 2.1626794258373207e-06, |
| "loss": 0.056909359991550446, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05856, |
| "step": 825, |
| "tokens/total": 3379200, |
| "tokens/train_per_sec_per_gpu": 20.35, |
| "tokens/trainable": 241974 |
| }, |
| { |
| "epoch": 3.9333333333333336, |
| "grad_norm": 0.4744802713394165, |
| "learning_rate": 2.15311004784689e-06, |
| "loss": 0.06698750704526901, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06928, |
| "step": 826, |
| "tokens/total": 3383296, |
| "tokens/train_per_sec_per_gpu": 18.28, |
| "tokens/trainable": 242326 |
| }, |
| { |
| "epoch": 3.9380952380952383, |
| "grad_norm": 0.9668700098991394, |
| "learning_rate": 2.1435406698564595e-06, |
| "loss": 0.1610293686389923, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.17472, |
| "step": 827, |
| "tokens/total": 3387392, |
| "tokens/train_per_sec_per_gpu": 14.63, |
| "tokens/trainable": 242623 |
| }, |
| { |
| "epoch": 3.942857142857143, |
| "grad_norm": 0.7834397554397583, |
| "learning_rate": 2.133971291866029e-06, |
| "loss": 0.0918939858675003, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09625, |
| "step": 828, |
| "tokens/total": 3391488, |
| "tokens/train_per_sec_per_gpu": 10.88, |
| "tokens/trainable": 242903 |
| }, |
| { |
| "epoch": 3.947619047619048, |
| "grad_norm": 0.6493245959281921, |
| "learning_rate": 2.1244019138755983e-06, |
| "loss": 0.06912299990653992, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07157, |
| "step": 829, |
| "tokens/total": 3395584, |
| "tokens/train_per_sec_per_gpu": 15.28, |
| "tokens/trainable": 243212 |
| }, |
| { |
| "epoch": 3.9523809523809526, |
| "grad_norm": 0.7439002394676208, |
| "learning_rate": 2.1148325358851677e-06, |
| "loss": 0.09127552807331085, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09557, |
| "step": 830, |
| "tokens/total": 3399680, |
| "tokens/train_per_sec_per_gpu": 9.96, |
| "tokens/trainable": 243543 |
| }, |
| { |
| "epoch": 3.9571428571428573, |
| "grad_norm": 1.073249101638794, |
| "learning_rate": 2.105263157894737e-06, |
| "loss": 0.07269235700368881, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0754, |
| "step": 831, |
| "tokens/total": 3403776, |
| "tokens/train_per_sec_per_gpu": 13.87, |
| "tokens/trainable": 243810 |
| }, |
| { |
| "epoch": 3.961904761904762, |
| "grad_norm": 0.914975643157959, |
| "learning_rate": 2.0956937799043064e-06, |
| "loss": 0.09338957071304321, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09789, |
| "step": 832, |
| "tokens/total": 3407872, |
| "tokens/train_per_sec_per_gpu": 22.28, |
| "tokens/trainable": 244184 |
| }, |
| { |
| "epoch": 3.966666666666667, |
| "grad_norm": 0.7717406153678894, |
| "learning_rate": 2.086124401913876e-06, |
| "loss": 0.08632536977529526, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09016, |
| "step": 833, |
| "tokens/total": 3411968, |
| "tokens/train_per_sec_per_gpu": 10.4, |
| "tokens/trainable": 244486 |
| }, |
| { |
| "epoch": 3.9714285714285715, |
| "grad_norm": 0.8154407739639282, |
| "learning_rate": 2.0765550239234452e-06, |
| "loss": 0.09252412617206573, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09694, |
| "step": 834, |
| "tokens/total": 3416064, |
| "tokens/train_per_sec_per_gpu": 15.65, |
| "tokens/trainable": 244821 |
| }, |
| { |
| "epoch": 3.9761904761904763, |
| "grad_norm": 0.7648400664329529, |
| "learning_rate": 2.0669856459330146e-06, |
| "loss": 0.06714729964733124, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06945, |
| "step": 835, |
| "tokens/total": 3420160, |
| "tokens/train_per_sec_per_gpu": 10.48, |
| "tokens/trainable": 245138 |
| }, |
| { |
| "epoch": 3.980952380952381, |
| "grad_norm": 0.6725287437438965, |
| "learning_rate": 2.057416267942584e-06, |
| "loss": 0.0852859616279602, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08903, |
| "step": 836, |
| "tokens/total": 3424256, |
| "tokens/train_per_sec_per_gpu": 16.19, |
| "tokens/trainable": 245447 |
| }, |
| { |
| "epoch": 3.9857142857142858, |
| "grad_norm": 0.5311653017997742, |
| "learning_rate": 2.0478468899521534e-06, |
| "loss": 0.05240347981452942, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0538, |
| "step": 837, |
| "tokens/total": 3428352, |
| "tokens/train_per_sec_per_gpu": 11.5, |
| "tokens/trainable": 245702 |
| }, |
| { |
| "epoch": 3.9904761904761905, |
| "grad_norm": 1.003235936164856, |
| "learning_rate": 2.0382775119617228e-06, |
| "loss": 0.09178285300731659, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09613, |
| "step": 838, |
| "tokens/total": 3432448, |
| "tokens/train_per_sec_per_gpu": 9.86, |
| "tokens/trainable": 245989 |
| }, |
| { |
| "epoch": 3.9952380952380953, |
| "grad_norm": 0.8917669653892517, |
| "learning_rate": 2.0287081339712917e-06, |
| "loss": 0.08628636598587036, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09012, |
| "step": 839, |
| "tokens/total": 3436544, |
| "tokens/train_per_sec_per_gpu": 11.31, |
| "tokens/trainable": 246249 |
| }, |
| { |
| "epoch": 4.0, |
| "grad_norm": 1.0130497217178345, |
| "learning_rate": 2.0191387559808616e-06, |
| "loss": 0.08156722784042358, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08499, |
| "step": 840, |
| "tokens/total": 3440640, |
| "tokens/train_per_sec_per_gpu": 12.54, |
| "tokens/trainable": 246513 |
| }, |
| { |
| "epoch": 4.0, |
| "eval_loss": 0.07883982360363007, |
| "eval_ppl": 1.08203, |
| "eval_runtime": 10.4308, |
| "eval_samples_per_second": 8.532, |
| "eval_steps_per_second": 2.205, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 35.59, |
| "memory/max_allocated (GiB)": 35.59, |
| "step": 840 |
| }, |
| { |
| "epoch": 4.004761904761905, |
| "grad_norm": 0.9490176439285278, |
| "learning_rate": 2.0095693779904305e-06, |
| "loss": 0.0749816969037056, |
| "memory/device_reserved (GiB)": 39.14, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07786, |
| "step": 841, |
| "tokens/total": 3444736, |
| "tokens/train_per_sec_per_gpu": 9.55, |
| "tokens/trainable": 246784 |
| }, |
| { |
| "epoch": 4.0095238095238095, |
| "grad_norm": 0.32807204127311707, |
| "learning_rate": 2.0000000000000003e-06, |
| "loss": 0.030588284134864807, |
| "memory/device_reserved (GiB)": 39.15, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03106, |
| "step": 842, |
| "tokens/total": 3448832, |
| "tokens/train_per_sec_per_gpu": 12.72, |
| "tokens/trainable": 247043 |
| }, |
| { |
| "epoch": 4.014285714285714, |
| "grad_norm": 1.0914336442947388, |
| "learning_rate": 1.9904306220095697e-06, |
| "loss": 0.1014956682920456, |
| "memory/device_reserved (GiB)": 39.15, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10683, |
| "step": 843, |
| "tokens/total": 3452928, |
| "tokens/train_per_sec_per_gpu": 11.84, |
| "tokens/trainable": 247297 |
| }, |
| { |
| "epoch": 4.019047619047619, |
| "grad_norm": 0.5259432792663574, |
| "learning_rate": 1.9808612440191387e-06, |
| "loss": 0.04887865483760834, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05009, |
| "step": 844, |
| "tokens/total": 3457024, |
| "tokens/train_per_sec_per_gpu": 15.25, |
| "tokens/trainable": 247594 |
| }, |
| { |
| "epoch": 4.023809523809524, |
| "grad_norm": 0.47249555587768555, |
| "learning_rate": 1.9712918660287085e-06, |
| "loss": 0.029536675661802292, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02998, |
| "step": 845, |
| "tokens/total": 3461120, |
| "tokens/train_per_sec_per_gpu": 9.89, |
| "tokens/trainable": 247848 |
| }, |
| { |
| "epoch": 4.0285714285714285, |
| "grad_norm": 0.6428244709968567, |
| "learning_rate": 1.9617224880382775e-06, |
| "loss": 0.05999932438135147, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06184, |
| "step": 846, |
| "tokens/total": 3465216, |
| "tokens/train_per_sec_per_gpu": 11.52, |
| "tokens/trainable": 248131 |
| }, |
| { |
| "epoch": 4.033333333333333, |
| "grad_norm": 0.6530048847198486, |
| "learning_rate": 1.952153110047847e-06, |
| "loss": 0.05361972004175186, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05508, |
| "step": 847, |
| "tokens/total": 3469312, |
| "tokens/train_per_sec_per_gpu": 13.34, |
| "tokens/trainable": 248414 |
| }, |
| { |
| "epoch": 4.038095238095238, |
| "grad_norm": 0.6523117423057556, |
| "learning_rate": 1.9425837320574167e-06, |
| "loss": 0.07807442545890808, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0812, |
| "step": 848, |
| "tokens/total": 3473408, |
| "tokens/train_per_sec_per_gpu": 15.01, |
| "tokens/trainable": 248740 |
| }, |
| { |
| "epoch": 4.042857142857143, |
| "grad_norm": 0.5848574638366699, |
| "learning_rate": 1.9330143540669856e-06, |
| "loss": 0.05901365354657173, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06079, |
| "step": 849, |
| "tokens/total": 3477504, |
| "tokens/train_per_sec_per_gpu": 14.06, |
| "tokens/trainable": 248990 |
| }, |
| { |
| "epoch": 4.0476190476190474, |
| "grad_norm": 0.6108696460723877, |
| "learning_rate": 1.9234449760765554e-06, |
| "loss": 0.05660616606473923, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05824, |
| "step": 850, |
| "tokens/total": 3481600, |
| "tokens/train_per_sec_per_gpu": 11.56, |
| "tokens/trainable": 249206 |
| }, |
| { |
| "epoch": 4.052380952380952, |
| "grad_norm": 0.7095998525619507, |
| "learning_rate": 1.9138755980861244e-06, |
| "loss": 0.06122968718409538, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06314, |
| "step": 851, |
| "tokens/total": 3485696, |
| "tokens/train_per_sec_per_gpu": 12.72, |
| "tokens/trainable": 249488 |
| }, |
| { |
| "epoch": 4.057142857142857, |
| "grad_norm": 0.7885696291923523, |
| "learning_rate": 1.9043062200956938e-06, |
| "loss": 0.061911046504974365, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06387, |
| "step": 852, |
| "tokens/total": 3489792, |
| "tokens/train_per_sec_per_gpu": 11.61, |
| "tokens/trainable": 249803 |
| }, |
| { |
| "epoch": 4.061904761904762, |
| "grad_norm": 0.5044105648994446, |
| "learning_rate": 1.8947368421052634e-06, |
| "loss": 0.05257502943277359, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05398, |
| "step": 853, |
| "tokens/total": 3493888, |
| "tokens/train_per_sec_per_gpu": 17.11, |
| "tokens/trainable": 250129 |
| }, |
| { |
| "epoch": 4.066666666666666, |
| "grad_norm": 0.9425777196884155, |
| "learning_rate": 1.8851674641148326e-06, |
| "loss": 0.10026482492685318, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10546, |
| "step": 854, |
| "tokens/total": 3497984, |
| "tokens/train_per_sec_per_gpu": 12.27, |
| "tokens/trainable": 250405 |
| }, |
| { |
| "epoch": 4.071428571428571, |
| "grad_norm": 0.9110622406005859, |
| "learning_rate": 1.8755980861244022e-06, |
| "loss": 0.11010429263114929, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11639, |
| "step": 855, |
| "tokens/total": 3502080, |
| "tokens/train_per_sec_per_gpu": 16.15, |
| "tokens/trainable": 250764 |
| }, |
| { |
| "epoch": 4.076190476190476, |
| "grad_norm": 1.1408833265304565, |
| "learning_rate": 1.8660287081339716e-06, |
| "loss": 0.08574778586626053, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.08953, |
| "step": 856, |
| "tokens/total": 3506176, |
| "tokens/train_per_sec_per_gpu": 16.3, |
| "tokens/trainable": 251035 |
| }, |
| { |
| "epoch": 4.080952380952381, |
| "grad_norm": 0.5738747715950012, |
| "learning_rate": 1.8564593301435407e-06, |
| "loss": 0.040827590972185135, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04167, |
| "step": 857, |
| "tokens/total": 3510272, |
| "tokens/train_per_sec_per_gpu": 14.79, |
| "tokens/trainable": 251397 |
| }, |
| { |
| "epoch": 4.085714285714285, |
| "grad_norm": 0.7939117550849915, |
| "learning_rate": 1.8468899521531103e-06, |
| "loss": 0.08306203782558441, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08661, |
| "step": 858, |
| "tokens/total": 3514368, |
| "tokens/train_per_sec_per_gpu": 13.97, |
| "tokens/trainable": 251738 |
| }, |
| { |
| "epoch": 4.09047619047619, |
| "grad_norm": 1.3952772617340088, |
| "learning_rate": 1.8373205741626795e-06, |
| "loss": 0.0917973592877388, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09614, |
| "step": 859, |
| "tokens/total": 3518464, |
| "tokens/train_per_sec_per_gpu": 14.95, |
| "tokens/trainable": 252071 |
| }, |
| { |
| "epoch": 4.095238095238095, |
| "grad_norm": 0.7479850053787231, |
| "learning_rate": 1.827751196172249e-06, |
| "loss": 0.06424461305141449, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06635, |
| "step": 860, |
| "tokens/total": 3522560, |
| "tokens/train_per_sec_per_gpu": 16.37, |
| "tokens/trainable": 252407 |
| }, |
| { |
| "epoch": 4.1, |
| "grad_norm": 0.9506804347038269, |
| "learning_rate": 1.8181818181818183e-06, |
| "loss": 0.08265562355518341, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08617, |
| "step": 861, |
| "tokens/total": 3526656, |
| "tokens/train_per_sec_per_gpu": 21.86, |
| "tokens/trainable": 252782 |
| }, |
| { |
| "epoch": 4.104761904761904, |
| "grad_norm": 0.7788168787956238, |
| "learning_rate": 1.8086124401913877e-06, |
| "loss": 0.04435598850250244, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04535, |
| "step": 862, |
| "tokens/total": 3530752, |
| "tokens/train_per_sec_per_gpu": 10.79, |
| "tokens/trainable": 253093 |
| }, |
| { |
| "epoch": 4.109523809523809, |
| "grad_norm": 0.2663736045360565, |
| "learning_rate": 1.7990430622009573e-06, |
| "loss": 0.026464153081178665, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.02682, |
| "step": 863, |
| "tokens/total": 3534848, |
| "tokens/train_per_sec_per_gpu": 11.88, |
| "tokens/trainable": 253415 |
| }, |
| { |
| "epoch": 4.114285714285714, |
| "grad_norm": 0.85882568359375, |
| "learning_rate": 1.7894736842105265e-06, |
| "loss": 0.07664963603019714, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07966, |
| "step": 864, |
| "tokens/total": 3538944, |
| "tokens/train_per_sec_per_gpu": 10.42, |
| "tokens/trainable": 253649 |
| }, |
| { |
| "epoch": 4.119047619047619, |
| "grad_norm": 0.6445422768592834, |
| "learning_rate": 1.7799043062200958e-06, |
| "loss": 0.0744195431470871, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07726, |
| "step": 865, |
| "tokens/total": 3543040, |
| "tokens/train_per_sec_per_gpu": 15.61, |
| "tokens/trainable": 253994 |
| }, |
| { |
| "epoch": 4.123809523809523, |
| "grad_norm": 1.065576195716858, |
| "learning_rate": 1.7703349282296652e-06, |
| "loss": 0.15491408109664917, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.16756, |
| "step": 866, |
| "tokens/total": 3547136, |
| "tokens/train_per_sec_per_gpu": 13.02, |
| "tokens/trainable": 254247 |
| }, |
| { |
| "epoch": 4.128571428571428, |
| "grad_norm": 0.4529033601284027, |
| "learning_rate": 1.7607655502392346e-06, |
| "loss": 0.034318629652261734, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03491, |
| "step": 867, |
| "tokens/total": 3551232, |
| "tokens/train_per_sec_per_gpu": 11.91, |
| "tokens/trainable": 254502 |
| }, |
| { |
| "epoch": 4.133333333333334, |
| "grad_norm": 0.7004088163375854, |
| "learning_rate": 1.7511961722488038e-06, |
| "loss": 0.08494338393211365, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08866, |
| "step": 868, |
| "tokens/total": 3555328, |
| "tokens/train_per_sec_per_gpu": 15.28, |
| "tokens/trainable": 254819 |
| }, |
| { |
| "epoch": 4.1380952380952385, |
| "grad_norm": 0.5470396876335144, |
| "learning_rate": 1.7416267942583734e-06, |
| "loss": 0.049904145300388336, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05117, |
| "step": 869, |
| "tokens/total": 3559424, |
| "tokens/train_per_sec_per_gpu": 13.28, |
| "tokens/trainable": 255094 |
| }, |
| { |
| "epoch": 4.142857142857143, |
| "grad_norm": 0.5020394921302795, |
| "learning_rate": 1.7320574162679426e-06, |
| "loss": 0.05947474017739296, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06128, |
| "step": 870, |
| "tokens/total": 3563520, |
| "tokens/train_per_sec_per_gpu": 15.45, |
| "tokens/trainable": 255414 |
| }, |
| { |
| "epoch": 4.147619047619048, |
| "grad_norm": 0.5141638517379761, |
| "learning_rate": 1.7224880382775122e-06, |
| "loss": 0.02415190264582634, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.02445, |
| "step": 871, |
| "tokens/total": 3567616, |
| "tokens/train_per_sec_per_gpu": 10.97, |
| "tokens/trainable": 255632 |
| }, |
| { |
| "epoch": 4.152380952380953, |
| "grad_norm": 1.4106786251068115, |
| "learning_rate": 1.7129186602870816e-06, |
| "loss": 0.10826382040977478, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11434, |
| "step": 872, |
| "tokens/total": 3571712, |
| "tokens/train_per_sec_per_gpu": 12.68, |
| "tokens/trainable": 255880 |
| }, |
| { |
| "epoch": 4.1571428571428575, |
| "grad_norm": 0.5615010261535645, |
| "learning_rate": 1.7033492822966507e-06, |
| "loss": 0.03849765658378601, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03925, |
| "step": 873, |
| "tokens/total": 3575808, |
| "tokens/train_per_sec_per_gpu": 14.63, |
| "tokens/trainable": 256174 |
| }, |
| { |
| "epoch": 4.161904761904762, |
| "grad_norm": 0.6795538067817688, |
| "learning_rate": 1.6937799043062203e-06, |
| "loss": 0.04768747463822365, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04884, |
| "step": 874, |
| "tokens/total": 3579904, |
| "tokens/train_per_sec_per_gpu": 25.87, |
| "tokens/trainable": 256568 |
| }, |
| { |
| "epoch": 4.166666666666667, |
| "grad_norm": 0.598505437374115, |
| "learning_rate": 1.6842105263157895e-06, |
| "loss": 0.05449119955301285, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.056, |
| "step": 875, |
| "tokens/total": 3584000, |
| "tokens/train_per_sec_per_gpu": 16.03, |
| "tokens/trainable": 256927 |
| }, |
| { |
| "epoch": 4.171428571428572, |
| "grad_norm": 0.8535393476486206, |
| "learning_rate": 1.6746411483253591e-06, |
| "loss": 0.07312190532684326, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07586, |
| "step": 876, |
| "tokens/total": 3588096, |
| "tokens/train_per_sec_per_gpu": 16.39, |
| "tokens/trainable": 257229 |
| }, |
| { |
| "epoch": 4.1761904761904765, |
| "grad_norm": 0.6472470164299011, |
| "learning_rate": 1.6650717703349283e-06, |
| "loss": 0.0614137165248394, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06334, |
| "step": 877, |
| "tokens/total": 3592192, |
| "tokens/train_per_sec_per_gpu": 11.11, |
| "tokens/trainable": 257506 |
| }, |
| { |
| "epoch": 4.180952380952381, |
| "grad_norm": 0.7363004684448242, |
| "learning_rate": 1.6555023923444977e-06, |
| "loss": 0.0493217408657074, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05056, |
| "step": 878, |
| "tokens/total": 3596288, |
| "tokens/train_per_sec_per_gpu": 9.68, |
| "tokens/trainable": 257818 |
| }, |
| { |
| "epoch": 4.185714285714286, |
| "grad_norm": 1.0971989631652832, |
| "learning_rate": 1.6459330143540673e-06, |
| "loss": 0.09783400595188141, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10278, |
| "step": 879, |
| "tokens/total": 3600384, |
| "tokens/train_per_sec_per_gpu": 18.66, |
| "tokens/trainable": 258201 |
| }, |
| { |
| "epoch": 4.190476190476191, |
| "grad_norm": 0.6767682433128357, |
| "learning_rate": 1.6363636363636365e-06, |
| "loss": 0.06688456237316132, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06917, |
| "step": 880, |
| "tokens/total": 3604480, |
| "tokens/train_per_sec_per_gpu": 17.66, |
| "tokens/trainable": 258495 |
| }, |
| { |
| "epoch": 4.195238095238095, |
| "grad_norm": 0.7312933802604675, |
| "learning_rate": 1.6267942583732059e-06, |
| "loss": 0.07728984951972961, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08036, |
| "step": 881, |
| "tokens/total": 3608576, |
| "tokens/train_per_sec_per_gpu": 15.97, |
| "tokens/trainable": 258800 |
| }, |
| { |
| "epoch": 4.2, |
| "grad_norm": 1.075819969177246, |
| "learning_rate": 1.6172248803827752e-06, |
| "loss": 0.09135079383850098, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09565, |
| "step": 882, |
| "tokens/total": 3612672, |
| "tokens/train_per_sec_per_gpu": 14.7, |
| "tokens/trainable": 259046 |
| }, |
| { |
| "epoch": 4.204761904761905, |
| "grad_norm": 0.7385833859443665, |
| "learning_rate": 1.6076555023923446e-06, |
| "loss": 0.10531473159790039, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11106, |
| "step": 883, |
| "tokens/total": 3616768, |
| "tokens/train_per_sec_per_gpu": 9.51, |
| "tokens/trainable": 259334 |
| }, |
| { |
| "epoch": 4.20952380952381, |
| "grad_norm": 0.6874673962593079, |
| "learning_rate": 1.5980861244019142e-06, |
| "loss": 0.09048204869031906, |
| "memory/device_reserved (GiB)": 39.17, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0947, |
| "step": 884, |
| "tokens/total": 3620864, |
| "tokens/train_per_sec_per_gpu": 17.49, |
| "tokens/trainable": 259691 |
| }, |
| { |
| "epoch": 4.214285714285714, |
| "grad_norm": 0.7855772972106934, |
| "learning_rate": 1.5885167464114834e-06, |
| "loss": 0.07922257483005524, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.08245, |
| "step": 885, |
| "tokens/total": 3624960, |
| "tokens/train_per_sec_per_gpu": 17.44, |
| "tokens/trainable": 260005 |
| }, |
| { |
| "epoch": 4.219047619047619, |
| "grad_norm": 0.6006829738616943, |
| "learning_rate": 1.5789473684210526e-06, |
| "loss": 0.06699418276548386, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06929, |
| "step": 886, |
| "tokens/total": 3629056, |
| "tokens/train_per_sec_per_gpu": 11.27, |
| "tokens/trainable": 260299 |
| }, |
| { |
| "epoch": 4.223809523809524, |
| "grad_norm": 0.6273767948150635, |
| "learning_rate": 1.5693779904306222e-06, |
| "loss": 0.07145752012729645, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07407, |
| "step": 887, |
| "tokens/total": 3633152, |
| "tokens/train_per_sec_per_gpu": 14.54, |
| "tokens/trainable": 260544 |
| }, |
| { |
| "epoch": 4.228571428571429, |
| "grad_norm": 0.6834083795547485, |
| "learning_rate": 1.5598086124401916e-06, |
| "loss": 0.06794765591621399, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07031, |
| "step": 888, |
| "tokens/total": 3637248, |
| "tokens/train_per_sec_per_gpu": 9.42, |
| "tokens/trainable": 260814 |
| }, |
| { |
| "epoch": 4.233333333333333, |
| "grad_norm": 0.49679332971572876, |
| "learning_rate": 1.5502392344497607e-06, |
| "loss": 0.061248525977134705, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06316, |
| "step": 889, |
| "tokens/total": 3641344, |
| "tokens/train_per_sec_per_gpu": 14.16, |
| "tokens/trainable": 261143 |
| }, |
| { |
| "epoch": 4.238095238095238, |
| "grad_norm": 0.5640564560890198, |
| "learning_rate": 1.5406698564593303e-06, |
| "loss": 0.0383354052901268, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.03908, |
| "step": 890, |
| "tokens/total": 3645440, |
| "tokens/train_per_sec_per_gpu": 14.15, |
| "tokens/trainable": 261469 |
| }, |
| { |
| "epoch": 4.242857142857143, |
| "grad_norm": 1.1639114618301392, |
| "learning_rate": 1.5311004784688995e-06, |
| "loss": 0.07877490669488907, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08196, |
| "step": 891, |
| "tokens/total": 3649536, |
| "tokens/train_per_sec_per_gpu": 11.55, |
| "tokens/trainable": 261719 |
| }, |
| { |
| "epoch": 4.247619047619048, |
| "grad_norm": 0.8218087553977966, |
| "learning_rate": 1.5215311004784691e-06, |
| "loss": 0.06628608703613281, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06853, |
| "step": 892, |
| "tokens/total": 3653632, |
| "tokens/train_per_sec_per_gpu": 10.83, |
| "tokens/trainable": 261995 |
| }, |
| { |
| "epoch": 4.252380952380952, |
| "grad_norm": 0.7092288732528687, |
| "learning_rate": 1.5119617224880383e-06, |
| "loss": 0.05123090371489525, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05257, |
| "step": 893, |
| "tokens/total": 3657728, |
| "tokens/train_per_sec_per_gpu": 14.36, |
| "tokens/trainable": 262321 |
| }, |
| { |
| "epoch": 4.257142857142857, |
| "grad_norm": 1.2661248445510864, |
| "learning_rate": 1.5023923444976077e-06, |
| "loss": 0.10810350626707077, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11416, |
| "step": 894, |
| "tokens/total": 3661824, |
| "tokens/train_per_sec_per_gpu": 11.75, |
| "tokens/trainable": 262652 |
| }, |
| { |
| "epoch": 4.261904761904762, |
| "grad_norm": 0.9689322113990784, |
| "learning_rate": 1.4928229665071773e-06, |
| "loss": 0.07231342047452927, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07499, |
| "step": 895, |
| "tokens/total": 3665920, |
| "tokens/train_per_sec_per_gpu": 10.82, |
| "tokens/trainable": 262895 |
| }, |
| { |
| "epoch": 4.266666666666667, |
| "grad_norm": 0.8054395914077759, |
| "learning_rate": 1.4832535885167465e-06, |
| "loss": 0.05944337695837021, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06125, |
| "step": 896, |
| "tokens/total": 3670016, |
| "tokens/train_per_sec_per_gpu": 15.87, |
| "tokens/trainable": 263163 |
| }, |
| { |
| "epoch": 4.271428571428571, |
| "grad_norm": 1.0870224237442017, |
| "learning_rate": 1.4736842105263159e-06, |
| "loss": 0.10757910460233688, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11358, |
| "step": 897, |
| "tokens/total": 3674112, |
| "tokens/train_per_sec_per_gpu": 14.39, |
| "tokens/trainable": 263459 |
| }, |
| { |
| "epoch": 4.276190476190476, |
| "grad_norm": 0.6998184323310852, |
| "learning_rate": 1.4641148325358852e-06, |
| "loss": 0.08821848034858704, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09223, |
| "step": 898, |
| "tokens/total": 3678208, |
| "tokens/train_per_sec_per_gpu": 14.14, |
| "tokens/trainable": 263726 |
| }, |
| { |
| "epoch": 4.280952380952381, |
| "grad_norm": 0.673193097114563, |
| "learning_rate": 1.4545454545454546e-06, |
| "loss": 0.05898454040288925, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06076, |
| "step": 899, |
| "tokens/total": 3682304, |
| "tokens/train_per_sec_per_gpu": 12.75, |
| "tokens/trainable": 263995 |
| }, |
| { |
| "epoch": 4.285714285714286, |
| "grad_norm": 0.9073561429977417, |
| "learning_rate": 1.4449760765550242e-06, |
| "loss": 0.09241870045661926, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09682, |
| "step": 900, |
| "tokens/total": 3686400, |
| "tokens/train_per_sec_per_gpu": 11.53, |
| "tokens/trainable": 264285 |
| }, |
| { |
| "epoch": 4.29047619047619, |
| "grad_norm": 0.6585322022438049, |
| "learning_rate": 1.4354066985645934e-06, |
| "loss": 0.0438724160194397, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04485, |
| "step": 901, |
| "tokens/total": 3690496, |
| "tokens/train_per_sec_per_gpu": 14.18, |
| "tokens/trainable": 264565 |
| }, |
| { |
| "epoch": 4.295238095238095, |
| "grad_norm": 0.6179357767105103, |
| "learning_rate": 1.4258373205741626e-06, |
| "loss": 0.04411366209387779, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0451, |
| "step": 902, |
| "tokens/total": 3694592, |
| "tokens/train_per_sec_per_gpu": 12.02, |
| "tokens/trainable": 264839 |
| }, |
| { |
| "epoch": 4.3, |
| "grad_norm": 0.6685640215873718, |
| "learning_rate": 1.4162679425837322e-06, |
| "loss": 0.07034850120544434, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07288, |
| "step": 903, |
| "tokens/total": 3698688, |
| "tokens/train_per_sec_per_gpu": 9.98, |
| "tokens/trainable": 265127 |
| }, |
| { |
| "epoch": 4.304761904761905, |
| "grad_norm": 0.9312854409217834, |
| "learning_rate": 1.4066985645933016e-06, |
| "loss": 0.13462990522384644, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.14411, |
| "step": 904, |
| "tokens/total": 3702784, |
| "tokens/train_per_sec_per_gpu": 16.23, |
| "tokens/trainable": 265471 |
| }, |
| { |
| "epoch": 4.309523809523809, |
| "grad_norm": 0.6181421279907227, |
| "learning_rate": 1.397129186602871e-06, |
| "loss": 0.07443803548812866, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07728, |
| "step": 905, |
| "tokens/total": 3706880, |
| "tokens/train_per_sec_per_gpu": 10.44, |
| "tokens/trainable": 265766 |
| }, |
| { |
| "epoch": 4.314285714285714, |
| "grad_norm": 0.903626561164856, |
| "learning_rate": 1.3875598086124404e-06, |
| "loss": 0.0580969899892807, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05982, |
| "step": 906, |
| "tokens/total": 3710976, |
| "tokens/train_per_sec_per_gpu": 13.25, |
| "tokens/trainable": 266057 |
| }, |
| { |
| "epoch": 4.319047619047619, |
| "grad_norm": 0.698418140411377, |
| "learning_rate": 1.3779904306220095e-06, |
| "loss": 0.06764661520719528, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06999, |
| "step": 907, |
| "tokens/total": 3715072, |
| "tokens/train_per_sec_per_gpu": 14.48, |
| "tokens/trainable": 266366 |
| }, |
| { |
| "epoch": 4.3238095238095235, |
| "grad_norm": 0.9411168098449707, |
| "learning_rate": 1.3684210526315791e-06, |
| "loss": 0.06708519160747528, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06939, |
| "step": 908, |
| "tokens/total": 3719168, |
| "tokens/train_per_sec_per_gpu": 13.63, |
| "tokens/trainable": 266633 |
| }, |
| { |
| "epoch": 4.328571428571428, |
| "grad_norm": 0.3852144479751587, |
| "learning_rate": 1.3588516746411485e-06, |
| "loss": 0.050158947706222534, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05144, |
| "step": 909, |
| "tokens/total": 3723264, |
| "tokens/train_per_sec_per_gpu": 13.9, |
| "tokens/trainable": 266989 |
| }, |
| { |
| "epoch": 4.333333333333333, |
| "grad_norm": 0.6585800647735596, |
| "learning_rate": 1.3492822966507177e-06, |
| "loss": 0.0391804464161396, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.03996, |
| "step": 910, |
| "tokens/total": 3727360, |
| "tokens/train_per_sec_per_gpu": 16.01, |
| "tokens/trainable": 267283 |
| }, |
| { |
| "epoch": 4.338095238095238, |
| "grad_norm": 0.6668374538421631, |
| "learning_rate": 1.3397129186602873e-06, |
| "loss": 0.04831254854798317, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0495, |
| "step": 911, |
| "tokens/total": 3731456, |
| "tokens/train_per_sec_per_gpu": 10.3, |
| "tokens/trainable": 267532 |
| }, |
| { |
| "epoch": 4.3428571428571425, |
| "grad_norm": 0.5903038382530212, |
| "learning_rate": 1.3301435406698565e-06, |
| "loss": 0.044688403606414795, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0457, |
| "step": 912, |
| "tokens/total": 3735552, |
| "tokens/train_per_sec_per_gpu": 17.5, |
| "tokens/trainable": 267819 |
| }, |
| { |
| "epoch": 4.347619047619047, |
| "grad_norm": 0.6831850409507751, |
| "learning_rate": 1.320574162679426e-06, |
| "loss": 0.06540787220001221, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06759, |
| "step": 913, |
| "tokens/total": 3739648, |
| "tokens/train_per_sec_per_gpu": 13.11, |
| "tokens/trainable": 268172 |
| }, |
| { |
| "epoch": 4.352380952380952, |
| "grad_norm": 0.9196857810020447, |
| "learning_rate": 1.3110047846889953e-06, |
| "loss": 0.10975831747055054, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11601, |
| "step": 914, |
| "tokens/total": 3743744, |
| "tokens/train_per_sec_per_gpu": 13.28, |
| "tokens/trainable": 268477 |
| }, |
| { |
| "epoch": 4.357142857142857, |
| "grad_norm": 0.710890531539917, |
| "learning_rate": 1.3014354066985646e-06, |
| "loss": 0.06814423203468323, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07052, |
| "step": 915, |
| "tokens/total": 3747840, |
| "tokens/train_per_sec_per_gpu": 15.36, |
| "tokens/trainable": 268765 |
| }, |
| { |
| "epoch": 4.3619047619047615, |
| "grad_norm": 0.9449613690376282, |
| "learning_rate": 1.2918660287081342e-06, |
| "loss": 0.0915263220667839, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09585, |
| "step": 916, |
| "tokens/total": 3751936, |
| "tokens/train_per_sec_per_gpu": 15.97, |
| "tokens/trainable": 269061 |
| }, |
| { |
| "epoch": 4.366666666666666, |
| "grad_norm": 1.084816575050354, |
| "learning_rate": 1.2822966507177034e-06, |
| "loss": 0.11780587583780289, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12503, |
| "step": 917, |
| "tokens/total": 3756032, |
| "tokens/train_per_sec_per_gpu": 13.84, |
| "tokens/trainable": 269372 |
| }, |
| { |
| "epoch": 4.371428571428572, |
| "grad_norm": 0.7779732942581177, |
| "learning_rate": 1.2727272727272728e-06, |
| "loss": 0.05686868727207184, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05852, |
| "step": 918, |
| "tokens/total": 3760128, |
| "tokens/train_per_sec_per_gpu": 18.35, |
| "tokens/trainable": 269688 |
| }, |
| { |
| "epoch": 4.376190476190477, |
| "grad_norm": 0.9143884181976318, |
| "learning_rate": 1.2631578947368422e-06, |
| "loss": 0.09882701933383942, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.10388, |
| "step": 919, |
| "tokens/total": 3764224, |
| "tokens/train_per_sec_per_gpu": 9.97, |
| "tokens/trainable": 269928 |
| }, |
| { |
| "epoch": 4.380952380952381, |
| "grad_norm": 0.6440621018409729, |
| "learning_rate": 1.2535885167464116e-06, |
| "loss": 0.06907931715250015, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07152, |
| "step": 920, |
| "tokens/total": 3768320, |
| "tokens/train_per_sec_per_gpu": 8.54, |
| "tokens/trainable": 270210 |
| }, |
| { |
| "epoch": 4.385714285714286, |
| "grad_norm": 0.623343825340271, |
| "learning_rate": 1.244019138755981e-06, |
| "loss": 0.04787050187587738, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04903, |
| "step": 921, |
| "tokens/total": 3772416, |
| "tokens/train_per_sec_per_gpu": 18.3, |
| "tokens/trainable": 270518 |
| }, |
| { |
| "epoch": 4.390476190476191, |
| "grad_norm": 0.7467364072799683, |
| "learning_rate": 1.2344497607655504e-06, |
| "loss": 0.09561535716056824, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10034, |
| "step": 922, |
| "tokens/total": 3776512, |
| "tokens/train_per_sec_per_gpu": 12.71, |
| "tokens/trainable": 270840 |
| }, |
| { |
| "epoch": 4.395238095238096, |
| "grad_norm": 0.9190205931663513, |
| "learning_rate": 1.2248803827751197e-06, |
| "loss": 0.08941492438316345, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09353, |
| "step": 923, |
| "tokens/total": 3780608, |
| "tokens/train_per_sec_per_gpu": 8.22, |
| "tokens/trainable": 271119 |
| }, |
| { |
| "epoch": 4.4, |
| "grad_norm": 0.5284753441810608, |
| "learning_rate": 1.2153110047846891e-06, |
| "loss": 0.03509580343961716, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.03572, |
| "step": 924, |
| "tokens/total": 3784704, |
| "tokens/train_per_sec_per_gpu": 13.56, |
| "tokens/trainable": 271362 |
| }, |
| { |
| "epoch": 4.404761904761905, |
| "grad_norm": 0.5473824739456177, |
| "learning_rate": 1.2057416267942585e-06, |
| "loss": 0.05125536769628525, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05259, |
| "step": 925, |
| "tokens/total": 3788800, |
| "tokens/train_per_sec_per_gpu": 14.61, |
| "tokens/trainable": 271686 |
| }, |
| { |
| "epoch": 4.40952380952381, |
| "grad_norm": 1.2579742670059204, |
| "learning_rate": 1.196172248803828e-06, |
| "loss": 0.08550336956977844, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08927, |
| "step": 926, |
| "tokens/total": 3792896, |
| "tokens/train_per_sec_per_gpu": 10.65, |
| "tokens/trainable": 271921 |
| }, |
| { |
| "epoch": 4.414285714285715, |
| "grad_norm": 0.8119210600852966, |
| "learning_rate": 1.1866028708133973e-06, |
| "loss": 0.050820425152778625, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05213, |
| "step": 927, |
| "tokens/total": 3796992, |
| "tokens/train_per_sec_per_gpu": 13.34, |
| "tokens/trainable": 272202 |
| }, |
| { |
| "epoch": 4.419047619047619, |
| "grad_norm": 1.0827769041061401, |
| "learning_rate": 1.1770334928229667e-06, |
| "loss": 0.05972447246313095, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06154, |
| "step": 928, |
| "tokens/total": 3801088, |
| "tokens/train_per_sec_per_gpu": 8.32, |
| "tokens/trainable": 272383 |
| }, |
| { |
| "epoch": 4.423809523809524, |
| "grad_norm": 0.8947194218635559, |
| "learning_rate": 1.1674641148325359e-06, |
| "loss": 0.05553201586008072, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0571, |
| "step": 929, |
| "tokens/total": 3805184, |
| "tokens/train_per_sec_per_gpu": 19.59, |
| "tokens/trainable": 272752 |
| }, |
| { |
| "epoch": 4.428571428571429, |
| "grad_norm": 0.7170199155807495, |
| "learning_rate": 1.1578947368421053e-06, |
| "loss": 0.04184604808688164, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04273, |
| "step": 930, |
| "tokens/total": 3809280, |
| "tokens/train_per_sec_per_gpu": 8.74, |
| "tokens/trainable": 272976 |
| }, |
| { |
| "epoch": 4.433333333333334, |
| "grad_norm": 1.015012264251709, |
| "learning_rate": 1.1483253588516749e-06, |
| "loss": 0.0780593678355217, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08119, |
| "step": 931, |
| "tokens/total": 3813376, |
| "tokens/train_per_sec_per_gpu": 11.31, |
| "tokens/trainable": 273249 |
| }, |
| { |
| "epoch": 4.438095238095238, |
| "grad_norm": 1.1480745077133179, |
| "learning_rate": 1.1387559808612442e-06, |
| "loss": 0.0772227793931961, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08028, |
| "step": 932, |
| "tokens/total": 3817472, |
| "tokens/train_per_sec_per_gpu": 10.67, |
| "tokens/trainable": 273536 |
| }, |
| { |
| "epoch": 4.442857142857143, |
| "grad_norm": 0.8914026618003845, |
| "learning_rate": 1.1291866028708134e-06, |
| "loss": 0.07296346873044968, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07569, |
| "step": 933, |
| "tokens/total": 3821568, |
| "tokens/train_per_sec_per_gpu": 11.68, |
| "tokens/trainable": 273807 |
| }, |
| { |
| "epoch": 4.447619047619048, |
| "grad_norm": 0.8942100405693054, |
| "learning_rate": 1.1196172248803828e-06, |
| "loss": 0.08658169955015182, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09044, |
| "step": 934, |
| "tokens/total": 3825664, |
| "tokens/train_per_sec_per_gpu": 12.8, |
| "tokens/trainable": 274172 |
| }, |
| { |
| "epoch": 4.4523809523809526, |
| "grad_norm": 1.1007604598999023, |
| "learning_rate": 1.1100478468899522e-06, |
| "loss": 0.06864183396100998, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07105, |
| "step": 935, |
| "tokens/total": 3829760, |
| "tokens/train_per_sec_per_gpu": 15.13, |
| "tokens/trainable": 274444 |
| }, |
| { |
| "epoch": 4.457142857142857, |
| "grad_norm": 0.4325333833694458, |
| "learning_rate": 1.1004784688995216e-06, |
| "loss": 0.04357939213514328, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04454, |
| "step": 936, |
| "tokens/total": 3833856, |
| "tokens/train_per_sec_per_gpu": 15.79, |
| "tokens/trainable": 274747 |
| }, |
| { |
| "epoch": 4.461904761904762, |
| "grad_norm": 0.50846928358078, |
| "learning_rate": 1.090909090909091e-06, |
| "loss": 0.056326571851968765, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05794, |
| "step": 937, |
| "tokens/total": 3837952, |
| "tokens/train_per_sec_per_gpu": 18.0, |
| "tokens/trainable": 275170 |
| }, |
| { |
| "epoch": 4.466666666666667, |
| "grad_norm": 0.902732253074646, |
| "learning_rate": 1.0813397129186604e-06, |
| "loss": 0.08314196765422821, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0867, |
| "step": 938, |
| "tokens/total": 3842048, |
| "tokens/train_per_sec_per_gpu": 14.19, |
| "tokens/trainable": 275447 |
| }, |
| { |
| "epoch": 4.4714285714285715, |
| "grad_norm": 0.5584244132041931, |
| "learning_rate": 1.0717703349282298e-06, |
| "loss": 0.046742670238018036, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04785, |
| "step": 939, |
| "tokens/total": 3846144, |
| "tokens/train_per_sec_per_gpu": 14.61, |
| "tokens/trainable": 275764 |
| }, |
| { |
| "epoch": 4.476190476190476, |
| "grad_norm": 1.0716779232025146, |
| "learning_rate": 1.0622009569377991e-06, |
| "loss": 0.09255675226449966, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09698, |
| "step": 940, |
| "tokens/total": 3850240, |
| "tokens/train_per_sec_per_gpu": 15.53, |
| "tokens/trainable": 276084 |
| }, |
| { |
| "epoch": 4.480952380952381, |
| "grad_norm": 0.7877600193023682, |
| "learning_rate": 1.0526315789473685e-06, |
| "loss": 0.10746045410633087, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11345, |
| "step": 941, |
| "tokens/total": 3854336, |
| "tokens/train_per_sec_per_gpu": 8.73, |
| "tokens/trainable": 276350 |
| }, |
| { |
| "epoch": 4.485714285714286, |
| "grad_norm": 1.062591314315796, |
| "learning_rate": 1.043062200956938e-06, |
| "loss": 0.08260044455528259, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08611, |
| "step": 942, |
| "tokens/total": 3858432, |
| "tokens/train_per_sec_per_gpu": 10.62, |
| "tokens/trainable": 276609 |
| }, |
| { |
| "epoch": 4.4904761904761905, |
| "grad_norm": 0.6927638053894043, |
| "learning_rate": 1.0334928229665073e-06, |
| "loss": 0.0631960779428482, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06524, |
| "step": 943, |
| "tokens/total": 3862528, |
| "tokens/train_per_sec_per_gpu": 17.25, |
| "tokens/trainable": 276908 |
| }, |
| { |
| "epoch": 4.495238095238095, |
| "grad_norm": 0.8995664119720459, |
| "learning_rate": 1.0239234449760767e-06, |
| "loss": 0.0754055604338646, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07832, |
| "step": 944, |
| "tokens/total": 3866624, |
| "tokens/train_per_sec_per_gpu": 14.29, |
| "tokens/trainable": 277204 |
| }, |
| { |
| "epoch": 4.5, |
| "grad_norm": 0.8859362006187439, |
| "learning_rate": 1.0143540669856459e-06, |
| "loss": 0.055626533925533295, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0572, |
| "step": 945, |
| "tokens/total": 3870720, |
| "tokens/train_per_sec_per_gpu": 11.05, |
| "tokens/trainable": 277423 |
| }, |
| { |
| "epoch": 4.504761904761905, |
| "grad_norm": 0.8782649040222168, |
| "learning_rate": 1.0047846889952153e-06, |
| "loss": 0.09288754314184189, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09734, |
| "step": 946, |
| "tokens/total": 3874816, |
| "tokens/train_per_sec_per_gpu": 10.56, |
| "tokens/trainable": 277757 |
| }, |
| { |
| "epoch": 4.5095238095238095, |
| "grad_norm": 0.659194827079773, |
| "learning_rate": 9.952153110047849e-07, |
| "loss": 0.04952794313430786, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05077, |
| "step": 947, |
| "tokens/total": 3878912, |
| "tokens/train_per_sec_per_gpu": 8.6, |
| "tokens/trainable": 278007 |
| }, |
| { |
| "epoch": 4.514285714285714, |
| "grad_norm": 1.2036082744598389, |
| "learning_rate": 9.856459330143542e-07, |
| "loss": 0.04865410178899765, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04986, |
| "step": 948, |
| "tokens/total": 3883008, |
| "tokens/train_per_sec_per_gpu": 13.01, |
| "tokens/trainable": 278237 |
| }, |
| { |
| "epoch": 4.519047619047619, |
| "grad_norm": 1.022188663482666, |
| "learning_rate": 9.760765550239234e-07, |
| "loss": 0.047886140644550323, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04905, |
| "step": 949, |
| "tokens/total": 3887104, |
| "tokens/train_per_sec_per_gpu": 13.16, |
| "tokens/trainable": 278478 |
| }, |
| { |
| "epoch": 4.523809523809524, |
| "grad_norm": 0.8995653390884399, |
| "learning_rate": 9.665071770334928e-07, |
| "loss": 0.07380396127700806, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0766, |
| "step": 950, |
| "tokens/total": 3891200, |
| "tokens/train_per_sec_per_gpu": 12.28, |
| "tokens/trainable": 278724 |
| }, |
| { |
| "epoch": 4.5285714285714285, |
| "grad_norm": 0.575376033782959, |
| "learning_rate": 9.569377990430622e-07, |
| "loss": 0.06823279708623886, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07061, |
| "step": 951, |
| "tokens/total": 3895296, |
| "tokens/train_per_sec_per_gpu": 15.52, |
| "tokens/trainable": 279033 |
| }, |
| { |
| "epoch": 4.533333333333333, |
| "grad_norm": 0.8048670887947083, |
| "learning_rate": 9.473684210526317e-07, |
| "loss": 0.06816083192825317, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07054, |
| "step": 952, |
| "tokens/total": 3899392, |
| "tokens/train_per_sec_per_gpu": 12.93, |
| "tokens/trainable": 279375 |
| }, |
| { |
| "epoch": 4.538095238095238, |
| "grad_norm": 0.632183849811554, |
| "learning_rate": 9.377990430622011e-07, |
| "loss": 0.04872383549809456, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04993, |
| "step": 953, |
| "tokens/total": 3903488, |
| "tokens/train_per_sec_per_gpu": 16.35, |
| "tokens/trainable": 279690 |
| }, |
| { |
| "epoch": 4.542857142857143, |
| "grad_norm": 0.726702094078064, |
| "learning_rate": 9.282296650717704e-07, |
| "loss": 0.07586966454982758, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07882, |
| "step": 954, |
| "tokens/total": 3907584, |
| "tokens/train_per_sec_per_gpu": 13.98, |
| "tokens/trainable": 279979 |
| }, |
| { |
| "epoch": 4.5476190476190474, |
| "grad_norm": 0.8905826807022095, |
| "learning_rate": 9.186602870813398e-07, |
| "loss": 0.09619905799627304, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10098, |
| "step": 955, |
| "tokens/total": 3911680, |
| "tokens/train_per_sec_per_gpu": 12.9, |
| "tokens/trainable": 280347 |
| }, |
| { |
| "epoch": 4.552380952380952, |
| "grad_norm": 0.7705732583999634, |
| "learning_rate": 9.090909090909091e-07, |
| "loss": 0.10297220945358276, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10846, |
| "step": 956, |
| "tokens/total": 3915776, |
| "tokens/train_per_sec_per_gpu": 10.51, |
| "tokens/trainable": 280613 |
| }, |
| { |
| "epoch": 4.557142857142857, |
| "grad_norm": 0.5715165138244629, |
| "learning_rate": 8.995215311004786e-07, |
| "loss": 0.04614497721195221, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04723, |
| "step": 957, |
| "tokens/total": 3919872, |
| "tokens/train_per_sec_per_gpu": 12.49, |
| "tokens/trainable": 280909 |
| }, |
| { |
| "epoch": 4.561904761904762, |
| "grad_norm": 0.9987708330154419, |
| "learning_rate": 8.899521531100479e-07, |
| "loss": 0.11006547510623932, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.11635, |
| "step": 958, |
| "tokens/total": 3923968, |
| "tokens/train_per_sec_per_gpu": 17.97, |
| "tokens/trainable": 281231 |
| }, |
| { |
| "epoch": 4.566666666666666, |
| "grad_norm": 0.42405691742897034, |
| "learning_rate": 8.803827751196173e-07, |
| "loss": 0.03897266089916229, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.03974, |
| "step": 959, |
| "tokens/total": 3928064, |
| "tokens/train_per_sec_per_gpu": 11.36, |
| "tokens/trainable": 281480 |
| }, |
| { |
| "epoch": 4.571428571428571, |
| "grad_norm": 1.079932451248169, |
| "learning_rate": 8.708133971291867e-07, |
| "loss": 0.09699150174856186, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10185, |
| "step": 960, |
| "tokens/total": 3932160, |
| "tokens/train_per_sec_per_gpu": 21.27, |
| "tokens/trainable": 281819 |
| }, |
| { |
| "epoch": 4.576190476190476, |
| "grad_norm": 0.5439162254333496, |
| "learning_rate": 8.612440191387561e-07, |
| "loss": 0.0659247413277626, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06815, |
| "step": 961, |
| "tokens/total": 3936256, |
| "tokens/train_per_sec_per_gpu": 13.75, |
| "tokens/trainable": 282164 |
| }, |
| { |
| "epoch": 4.580952380952381, |
| "grad_norm": 0.8237618803977966, |
| "learning_rate": 8.516746411483254e-07, |
| "loss": 0.08178961277008057, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08523, |
| "step": 962, |
| "tokens/total": 3940352, |
| "tokens/train_per_sec_per_gpu": 9.45, |
| "tokens/trainable": 282383 |
| }, |
| { |
| "epoch": 4.585714285714285, |
| "grad_norm": 0.5932995080947876, |
| "learning_rate": 8.421052631578948e-07, |
| "loss": 0.0472625270485878, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0484, |
| "step": 963, |
| "tokens/total": 3944448, |
| "tokens/train_per_sec_per_gpu": 11.99, |
| "tokens/trainable": 282631 |
| }, |
| { |
| "epoch": 4.59047619047619, |
| "grad_norm": 0.6871050000190735, |
| "learning_rate": 8.325358851674641e-07, |
| "loss": 0.07640991359949112, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0794, |
| "step": 964, |
| "tokens/total": 3948544, |
| "tokens/train_per_sec_per_gpu": 12.59, |
| "tokens/trainable": 282857 |
| }, |
| { |
| "epoch": 4.595238095238095, |
| "grad_norm": 0.7261757850646973, |
| "learning_rate": 8.229665071770336e-07, |
| "loss": 0.05905777961015701, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06084, |
| "step": 965, |
| "tokens/total": 3952640, |
| "tokens/train_per_sec_per_gpu": 15.62, |
| "tokens/trainable": 283162 |
| }, |
| { |
| "epoch": 4.6, |
| "grad_norm": 0.897055447101593, |
| "learning_rate": 8.133971291866029e-07, |
| "loss": 0.0663570985198021, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06861, |
| "step": 966, |
| "tokens/total": 3956736, |
| "tokens/train_per_sec_per_gpu": 10.57, |
| "tokens/trainable": 283396 |
| }, |
| { |
| "epoch": 4.604761904761904, |
| "grad_norm": 0.5905396342277527, |
| "learning_rate": 8.038277511961723e-07, |
| "loss": 0.05556555464863777, |
| "memory/device_reserved (GiB)": 39.19, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05714, |
| "step": 967, |
| "tokens/total": 3960832, |
| "tokens/train_per_sec_per_gpu": 16.93, |
| "tokens/trainable": 283725 |
| }, |
| { |
| "epoch": 4.609523809523809, |
| "grad_norm": 0.4669351875782013, |
| "learning_rate": 7.942583732057417e-07, |
| "loss": 0.048831816762685776, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.57, |
| "memory/max_allocated (GiB)": 38.57, |
| "ppl": 1.05004, |
| "step": 968, |
| "tokens/total": 3964928, |
| "tokens/train_per_sec_per_gpu": 26.39, |
| "tokens/trainable": 284224 |
| }, |
| { |
| "epoch": 4.614285714285714, |
| "grad_norm": 0.5662763714790344, |
| "learning_rate": 7.846889952153111e-07, |
| "loss": 0.0546397939324379, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05616, |
| "step": 969, |
| "tokens/total": 3969024, |
| "tokens/train_per_sec_per_gpu": 18.62, |
| "tokens/trainable": 284553 |
| }, |
| { |
| "epoch": 4.619047619047619, |
| "grad_norm": 0.7985778450965881, |
| "learning_rate": 7.751196172248804e-07, |
| "loss": 0.0845763310790062, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08826, |
| "step": 970, |
| "tokens/total": 3973120, |
| "tokens/train_per_sec_per_gpu": 9.05, |
| "tokens/trainable": 284758 |
| }, |
| { |
| "epoch": 4.623809523809523, |
| "grad_norm": 0.6422544717788696, |
| "learning_rate": 7.655502392344498e-07, |
| "loss": 0.0869225338101387, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09081, |
| "step": 971, |
| "tokens/total": 3977216, |
| "tokens/train_per_sec_per_gpu": 19.71, |
| "tokens/trainable": 285169 |
| }, |
| { |
| "epoch": 4.628571428571428, |
| "grad_norm": 0.7870466113090515, |
| "learning_rate": 7.559808612440192e-07, |
| "loss": 0.09116926789283752, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09545, |
| "step": 972, |
| "tokens/total": 3981312, |
| "tokens/train_per_sec_per_gpu": 13.42, |
| "tokens/trainable": 285429 |
| }, |
| { |
| "epoch": 4.633333333333333, |
| "grad_norm": 0.5808945894241333, |
| "learning_rate": 7.464114832535886e-07, |
| "loss": 0.07630980759859085, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0793, |
| "step": 973, |
| "tokens/total": 3985408, |
| "tokens/train_per_sec_per_gpu": 17.25, |
| "tokens/trainable": 285864 |
| }, |
| { |
| "epoch": 4.638095238095238, |
| "grad_norm": 0.5704347491264343, |
| "learning_rate": 7.368421052631579e-07, |
| "loss": 0.06161660701036453, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06355, |
| "step": 974, |
| "tokens/total": 3989504, |
| "tokens/train_per_sec_per_gpu": 12.91, |
| "tokens/trainable": 286193 |
| }, |
| { |
| "epoch": 4.642857142857143, |
| "grad_norm": 0.4083561897277832, |
| "learning_rate": 7.272727272727273e-07, |
| "loss": 0.04932102560997009, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05056, |
| "step": 975, |
| "tokens/total": 3993600, |
| "tokens/train_per_sec_per_gpu": 17.5, |
| "tokens/trainable": 286572 |
| }, |
| { |
| "epoch": 4.647619047619048, |
| "grad_norm": 0.7787036895751953, |
| "learning_rate": 7.177033492822967e-07, |
| "loss": 0.0850168764591217, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08874, |
| "step": 976, |
| "tokens/total": 3997696, |
| "tokens/train_per_sec_per_gpu": 18.18, |
| "tokens/trainable": 286900 |
| }, |
| { |
| "epoch": 4.652380952380953, |
| "grad_norm": 0.7790368795394897, |
| "learning_rate": 7.081339712918661e-07, |
| "loss": 0.08561737090349197, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08939, |
| "step": 977, |
| "tokens/total": 4001792, |
| "tokens/train_per_sec_per_gpu": 10.6, |
| "tokens/trainable": 287213 |
| }, |
| { |
| "epoch": 4.6571428571428575, |
| "grad_norm": 0.725902795791626, |
| "learning_rate": 6.985645933014355e-07, |
| "loss": 0.04747258126735687, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04862, |
| "step": 978, |
| "tokens/total": 4005888, |
| "tokens/train_per_sec_per_gpu": 12.51, |
| "tokens/trainable": 287508 |
| }, |
| { |
| "epoch": 4.661904761904762, |
| "grad_norm": 0.5647050142288208, |
| "learning_rate": 6.889952153110048e-07, |
| "loss": 0.04467792436480522, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04569, |
| "step": 979, |
| "tokens/total": 4009984, |
| "tokens/train_per_sec_per_gpu": 9.26, |
| "tokens/trainable": 287738 |
| }, |
| { |
| "epoch": 4.666666666666667, |
| "grad_norm": 0.7799670696258545, |
| "learning_rate": 6.794258373205743e-07, |
| "loss": 0.06656090170145035, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06883, |
| "step": 980, |
| "tokens/total": 4014080, |
| "tokens/train_per_sec_per_gpu": 15.49, |
| "tokens/trainable": 288091 |
| }, |
| { |
| "epoch": 4.671428571428572, |
| "grad_norm": 0.8358564376831055, |
| "learning_rate": 6.698564593301436e-07, |
| "loss": 0.07533138990402222, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07824, |
| "step": 981, |
| "tokens/total": 4018176, |
| "tokens/train_per_sec_per_gpu": 28.76, |
| "tokens/trainable": 288533 |
| }, |
| { |
| "epoch": 4.6761904761904765, |
| "grad_norm": 0.6379777789115906, |
| "learning_rate": 6.60287081339713e-07, |
| "loss": 0.054609671235084534, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05613, |
| "step": 982, |
| "tokens/total": 4022272, |
| "tokens/train_per_sec_per_gpu": 18.17, |
| "tokens/trainable": 288821 |
| }, |
| { |
| "epoch": 4.680952380952381, |
| "grad_norm": 0.8028697371482849, |
| "learning_rate": 6.507177033492823e-07, |
| "loss": 0.0740557312965393, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07687, |
| "step": 983, |
| "tokens/total": 4026368, |
| "tokens/train_per_sec_per_gpu": 13.12, |
| "tokens/trainable": 289092 |
| }, |
| { |
| "epoch": 4.685714285714286, |
| "grad_norm": 0.9271824955940247, |
| "learning_rate": 6.411483253588517e-07, |
| "loss": 0.09886521100997925, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10392, |
| "step": 984, |
| "tokens/total": 4030464, |
| "tokens/train_per_sec_per_gpu": 10.96, |
| "tokens/trainable": 289321 |
| }, |
| { |
| "epoch": 4.690476190476191, |
| "grad_norm": 0.7451408505439758, |
| "learning_rate": 6.315789473684211e-07, |
| "loss": 0.04617952182888985, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04726, |
| "step": 985, |
| "tokens/total": 4034560, |
| "tokens/train_per_sec_per_gpu": 11.97, |
| "tokens/trainable": 289574 |
| }, |
| { |
| "epoch": 4.695238095238095, |
| "grad_norm": 0.8016191720962524, |
| "learning_rate": 6.220095693779905e-07, |
| "loss": 0.061083678156137466, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06299, |
| "step": 986, |
| "tokens/total": 4038656, |
| "tokens/train_per_sec_per_gpu": 10.8, |
| "tokens/trainable": 289835 |
| }, |
| { |
| "epoch": 4.7, |
| "grad_norm": 0.5138081312179565, |
| "learning_rate": 6.124401913875599e-07, |
| "loss": 0.04855319485068321, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04975, |
| "step": 987, |
| "tokens/total": 4042752, |
| "tokens/train_per_sec_per_gpu": 14.26, |
| "tokens/trainable": 290171 |
| }, |
| { |
| "epoch": 4.704761904761905, |
| "grad_norm": 0.35503482818603516, |
| "learning_rate": 6.028708133971293e-07, |
| "loss": 0.025968704372644424, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.02631, |
| "step": 988, |
| "tokens/total": 4046848, |
| "tokens/train_per_sec_per_gpu": 13.11, |
| "tokens/trainable": 290471 |
| }, |
| { |
| "epoch": 4.70952380952381, |
| "grad_norm": 1.100373387336731, |
| "learning_rate": 5.933014354066987e-07, |
| "loss": 0.05797182768583298, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05969, |
| "step": 989, |
| "tokens/total": 4050944, |
| "tokens/train_per_sec_per_gpu": 17.54, |
| "tokens/trainable": 290763 |
| }, |
| { |
| "epoch": 4.714285714285714, |
| "grad_norm": 0.679127037525177, |
| "learning_rate": 5.837320574162679e-07, |
| "loss": 0.041285596787929535, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04215, |
| "step": 990, |
| "tokens/total": 4055040, |
| "tokens/train_per_sec_per_gpu": 11.44, |
| "tokens/trainable": 291038 |
| }, |
| { |
| "epoch": 4.719047619047619, |
| "grad_norm": 0.6147889494895935, |
| "learning_rate": 5.741626794258374e-07, |
| "loss": 0.04121209308505058, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04207, |
| "step": 991, |
| "tokens/total": 4059136, |
| "tokens/train_per_sec_per_gpu": 12.09, |
| "tokens/trainable": 291278 |
| }, |
| { |
| "epoch": 4.723809523809524, |
| "grad_norm": 0.8105634450912476, |
| "learning_rate": 5.645933014354067e-07, |
| "loss": 0.06966257840394974, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07215, |
| "step": 992, |
| "tokens/total": 4063232, |
| "tokens/train_per_sec_per_gpu": 12.89, |
| "tokens/trainable": 291563 |
| }, |
| { |
| "epoch": 4.728571428571429, |
| "grad_norm": 0.5204731822013855, |
| "learning_rate": 5.550239234449761e-07, |
| "loss": 0.055574752390384674, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05715, |
| "step": 993, |
| "tokens/total": 4067328, |
| "tokens/train_per_sec_per_gpu": 11.17, |
| "tokens/trainable": 291921 |
| }, |
| { |
| "epoch": 4.733333333333333, |
| "grad_norm": 1.142088770866394, |
| "learning_rate": 5.454545454545455e-07, |
| "loss": 0.09287901967763901, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09733, |
| "step": 994, |
| "tokens/total": 4071424, |
| "tokens/train_per_sec_per_gpu": 13.94, |
| "tokens/trainable": 292187 |
| }, |
| { |
| "epoch": 4.738095238095238, |
| "grad_norm": 0.5879678726196289, |
| "learning_rate": 5.358851674641149e-07, |
| "loss": 0.04538918659090996, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04644, |
| "step": 995, |
| "tokens/total": 4075520, |
| "tokens/train_per_sec_per_gpu": 9.32, |
| "tokens/trainable": 292429 |
| }, |
| { |
| "epoch": 4.742857142857143, |
| "grad_norm": 0.6333562135696411, |
| "learning_rate": 5.263157894736843e-07, |
| "loss": 0.05799243599176407, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.05971, |
| "step": 996, |
| "tokens/total": 4079616, |
| "tokens/train_per_sec_per_gpu": 13.4, |
| "tokens/trainable": 292716 |
| }, |
| { |
| "epoch": 4.747619047619048, |
| "grad_norm": 0.4939824342727661, |
| "learning_rate": 5.167464114832537e-07, |
| "loss": 0.042676541954278946, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0436, |
| "step": 997, |
| "tokens/total": 4083712, |
| "tokens/train_per_sec_per_gpu": 12.45, |
| "tokens/trainable": 293041 |
| }, |
| { |
| "epoch": 4.752380952380952, |
| "grad_norm": 0.7977603673934937, |
| "learning_rate": 5.071770334928229e-07, |
| "loss": 0.06747094541788101, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0698, |
| "step": 998, |
| "tokens/total": 4087808, |
| "tokens/train_per_sec_per_gpu": 12.58, |
| "tokens/trainable": 293352 |
| }, |
| { |
| "epoch": 4.757142857142857, |
| "grad_norm": 0.7418789863586426, |
| "learning_rate": 4.976076555023924e-07, |
| "loss": 0.07924629002809525, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.08247, |
| "step": 999, |
| "tokens/total": 4091904, |
| "tokens/train_per_sec_per_gpu": 16.42, |
| "tokens/trainable": 293763 |
| }, |
| { |
| "epoch": 4.761904761904762, |
| "grad_norm": 0.8634079098701477, |
| "learning_rate": 4.880382775119617e-07, |
| "loss": 0.06551891565322876, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06771, |
| "step": 1000, |
| "tokens/total": 4096000, |
| "tokens/train_per_sec_per_gpu": 13.6, |
| "tokens/trainable": 294007 |
| }, |
| { |
| "epoch": 4.766666666666667, |
| "grad_norm": 0.9717041850090027, |
| "learning_rate": 4.784688995215311e-07, |
| "loss": 0.11543455719947815, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.12236, |
| "step": 1001, |
| "tokens/total": 4100096, |
| "tokens/train_per_sec_per_gpu": 10.13, |
| "tokens/trainable": 294235 |
| }, |
| { |
| "epoch": 4.771428571428571, |
| "grad_norm": 0.6794291138648987, |
| "learning_rate": 4.6889952153110054e-07, |
| "loss": 0.06426049023866653, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06637, |
| "step": 1002, |
| "tokens/total": 4104192, |
| "tokens/train_per_sec_per_gpu": 9.06, |
| "tokens/trainable": 294445 |
| }, |
| { |
| "epoch": 4.776190476190476, |
| "grad_norm": 0.9054945707321167, |
| "learning_rate": 4.593301435406699e-07, |
| "loss": 0.07165690511465073, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07429, |
| "step": 1003, |
| "tokens/total": 4108288, |
| "tokens/train_per_sec_per_gpu": 13.8, |
| "tokens/trainable": 294722 |
| }, |
| { |
| "epoch": 4.780952380952381, |
| "grad_norm": 0.8471719622612, |
| "learning_rate": 4.497607655502393e-07, |
| "loss": 0.0860835611820221, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.0899, |
| "step": 1004, |
| "tokens/total": 4112384, |
| "tokens/train_per_sec_per_gpu": 15.63, |
| "tokens/trainable": 295030 |
| }, |
| { |
| "epoch": 4.785714285714286, |
| "grad_norm": 0.4434499144554138, |
| "learning_rate": 4.4019138755980866e-07, |
| "loss": 0.036177873611450195, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.03684, |
| "step": 1005, |
| "tokens/total": 4116480, |
| "tokens/train_per_sec_per_gpu": 12.9, |
| "tokens/trainable": 295290 |
| }, |
| { |
| "epoch": 4.79047619047619, |
| "grad_norm": 0.6867081522941589, |
| "learning_rate": 4.3062200956937804e-07, |
| "loss": 0.07819761335849762, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08134, |
| "step": 1006, |
| "tokens/total": 4120576, |
| "tokens/train_per_sec_per_gpu": 14.01, |
| "tokens/trainable": 295572 |
| }, |
| { |
| "epoch": 4.795238095238095, |
| "grad_norm": 0.869951069355011, |
| "learning_rate": 4.210526315789474e-07, |
| "loss": 0.047713182866573334, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04887, |
| "step": 1007, |
| "tokens/total": 4124672, |
| "tokens/train_per_sec_per_gpu": 12.37, |
| "tokens/trainable": 295799 |
| }, |
| { |
| "epoch": 4.8, |
| "grad_norm": 0.4378795921802521, |
| "learning_rate": 4.114832535885168e-07, |
| "loss": 0.05229447782039642, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05369, |
| "step": 1008, |
| "tokens/total": 4128768, |
| "tokens/train_per_sec_per_gpu": 27.47, |
| "tokens/trainable": 296186 |
| }, |
| { |
| "epoch": 4.804761904761905, |
| "grad_norm": 0.7369760870933533, |
| "learning_rate": 4.0191387559808616e-07, |
| "loss": 0.044564686715602875, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04557, |
| "step": 1009, |
| "tokens/total": 4132864, |
| "tokens/train_per_sec_per_gpu": 9.89, |
| "tokens/trainable": 296474 |
| }, |
| { |
| "epoch": 4.809523809523809, |
| "grad_norm": 0.7173194289207458, |
| "learning_rate": 3.9234449760765555e-07, |
| "loss": 0.05833519622683525, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06007, |
| "step": 1010, |
| "tokens/total": 4136960, |
| "tokens/train_per_sec_per_gpu": 11.28, |
| "tokens/trainable": 296730 |
| }, |
| { |
| "epoch": 4.814285714285714, |
| "grad_norm": 0.5104710459709167, |
| "learning_rate": 3.827751196172249e-07, |
| "loss": 0.0309146698564291, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0314, |
| "step": 1011, |
| "tokens/total": 4141056, |
| "tokens/train_per_sec_per_gpu": 10.74, |
| "tokens/trainable": 296960 |
| }, |
| { |
| "epoch": 4.819047619047619, |
| "grad_norm": 0.835244357585907, |
| "learning_rate": 3.732057416267943e-07, |
| "loss": 0.06693878769874573, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06923, |
| "step": 1012, |
| "tokens/total": 4145152, |
| "tokens/train_per_sec_per_gpu": 11.61, |
| "tokens/trainable": 297187 |
| }, |
| { |
| "epoch": 4.8238095238095235, |
| "grad_norm": 0.8610392808914185, |
| "learning_rate": 3.6363636363636366e-07, |
| "loss": 0.05454925820231438, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05606, |
| "step": 1013, |
| "tokens/total": 4149248, |
| "tokens/train_per_sec_per_gpu": 14.69, |
| "tokens/trainable": 297538 |
| }, |
| { |
| "epoch": 4.828571428571428, |
| "grad_norm": 0.8053697943687439, |
| "learning_rate": 3.5406698564593305e-07, |
| "loss": 0.059003330767154694, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06078, |
| "step": 1014, |
| "tokens/total": 4153344, |
| "tokens/train_per_sec_per_gpu": 10.53, |
| "tokens/trainable": 297788 |
| }, |
| { |
| "epoch": 4.833333333333333, |
| "grad_norm": 0.8027690649032593, |
| "learning_rate": 3.444976076555024e-07, |
| "loss": 0.06386961042881012, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06595, |
| "step": 1015, |
| "tokens/total": 4157440, |
| "tokens/train_per_sec_per_gpu": 12.25, |
| "tokens/trainable": 298051 |
| }, |
| { |
| "epoch": 4.838095238095238, |
| "grad_norm": 0.5434281229972839, |
| "learning_rate": 3.349282296650718e-07, |
| "loss": 0.06257535517215729, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.06457, |
| "step": 1016, |
| "tokens/total": 4161536, |
| "tokens/train_per_sec_per_gpu": 18.94, |
| "tokens/trainable": 298439 |
| }, |
| { |
| "epoch": 4.8428571428571425, |
| "grad_norm": 0.8059958219528198, |
| "learning_rate": 3.2535885167464116e-07, |
| "loss": 0.080122210085392, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.08342, |
| "step": 1017, |
| "tokens/total": 4165632, |
| "tokens/train_per_sec_per_gpu": 15.14, |
| "tokens/trainable": 298717 |
| }, |
| { |
| "epoch": 4.847619047619047, |
| "grad_norm": 0.7690683603286743, |
| "learning_rate": 3.1578947368421055e-07, |
| "loss": 0.04988192766904831, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05115, |
| "step": 1018, |
| "tokens/total": 4169728, |
| "tokens/train_per_sec_per_gpu": 10.57, |
| "tokens/trainable": 298919 |
| }, |
| { |
| "epoch": 4.852380952380952, |
| "grad_norm": 1.0210976600646973, |
| "learning_rate": 3.0622009569377994e-07, |
| "loss": 0.07849578559398651, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.08166, |
| "step": 1019, |
| "tokens/total": 4173824, |
| "tokens/train_per_sec_per_gpu": 11.01, |
| "tokens/trainable": 299160 |
| }, |
| { |
| "epoch": 4.857142857142857, |
| "grad_norm": 0.9019124507904053, |
| "learning_rate": 2.966507177033493e-07, |
| "loss": 0.09448347240686417, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.09909, |
| "step": 1020, |
| "tokens/total": 4177920, |
| "tokens/train_per_sec_per_gpu": 16.98, |
| "tokens/trainable": 299460 |
| }, |
| { |
| "epoch": 4.861904761904762, |
| "grad_norm": 0.8207787275314331, |
| "learning_rate": 2.870813397129187e-07, |
| "loss": 0.049627967178821564, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05088, |
| "step": 1021, |
| "tokens/total": 4182016, |
| "tokens/train_per_sec_per_gpu": 11.28, |
| "tokens/trainable": 299738 |
| }, |
| { |
| "epoch": 4.866666666666667, |
| "grad_norm": 0.6081165671348572, |
| "learning_rate": 2.7751196172248805e-07, |
| "loss": 0.05352705717086792, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05499, |
| "step": 1022, |
| "tokens/total": 4186112, |
| "tokens/train_per_sec_per_gpu": 9.65, |
| "tokens/trainable": 300042 |
| }, |
| { |
| "epoch": 4.871428571428572, |
| "grad_norm": 1.1343626976013184, |
| "learning_rate": 2.6794258373205744e-07, |
| "loss": 0.06644172966480255, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0687, |
| "step": 1023, |
| "tokens/total": 4190208, |
| "tokens/train_per_sec_per_gpu": 13.68, |
| "tokens/trainable": 300285 |
| }, |
| { |
| "epoch": 4.876190476190477, |
| "grad_norm": 0.8928014636039734, |
| "learning_rate": 2.5837320574162683e-07, |
| "loss": 0.08010731637477875, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0834, |
| "step": 1024, |
| "tokens/total": 4194304, |
| "tokens/train_per_sec_per_gpu": 21.81, |
| "tokens/trainable": 300641 |
| }, |
| { |
| "epoch": 4.880952380952381, |
| "grad_norm": 1.156931757926941, |
| "learning_rate": 2.488038277511962e-07, |
| "loss": 0.11981767416000366, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.12729, |
| "step": 1025, |
| "tokens/total": 4198400, |
| "tokens/train_per_sec_per_gpu": 9.24, |
| "tokens/trainable": 300918 |
| }, |
| { |
| "epoch": 4.885714285714286, |
| "grad_norm": 0.6716881990432739, |
| "learning_rate": 2.3923444976076555e-07, |
| "loss": 0.06406470388174057, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06616, |
| "step": 1026, |
| "tokens/total": 4202496, |
| "tokens/train_per_sec_per_gpu": 11.3, |
| "tokens/trainable": 301204 |
| }, |
| { |
| "epoch": 4.890476190476191, |
| "grad_norm": 1.2062376737594604, |
| "learning_rate": 2.2966507177033494e-07, |
| "loss": 0.09758275747299194, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.1025, |
| "step": 1027, |
| "tokens/total": 4206592, |
| "tokens/train_per_sec_per_gpu": 15.1, |
| "tokens/trainable": 301487 |
| }, |
| { |
| "epoch": 4.895238095238096, |
| "grad_norm": 0.7098796367645264, |
| "learning_rate": 2.2009569377990433e-07, |
| "loss": 0.06214335933327675, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.56, |
| "memory/max_allocated (GiB)": 38.56, |
| "ppl": 1.06411, |
| "step": 1028, |
| "tokens/total": 4210688, |
| "tokens/train_per_sec_per_gpu": 12.37, |
| "tokens/trainable": 301759 |
| }, |
| { |
| "epoch": 4.9, |
| "grad_norm": 1.2185698747634888, |
| "learning_rate": 2.105263157894737e-07, |
| "loss": 0.04552961885929108, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04658, |
| "step": 1029, |
| "tokens/total": 4214784, |
| "tokens/train_per_sec_per_gpu": 13.94, |
| "tokens/trainable": 302032 |
| }, |
| { |
| "epoch": 4.904761904761905, |
| "grad_norm": 0.7723051309585571, |
| "learning_rate": 2.0095693779904308e-07, |
| "loss": 0.07359997183084488, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07638, |
| "step": 1030, |
| "tokens/total": 4218880, |
| "tokens/train_per_sec_per_gpu": 9.81, |
| "tokens/trainable": 302240 |
| }, |
| { |
| "epoch": 4.90952380952381, |
| "grad_norm": 0.46518200635910034, |
| "learning_rate": 1.9138755980861244e-07, |
| "loss": 0.04536569491028786, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04641, |
| "step": 1031, |
| "tokens/total": 4222976, |
| "tokens/train_per_sec_per_gpu": 11.09, |
| "tokens/trainable": 302490 |
| }, |
| { |
| "epoch": 4.914285714285715, |
| "grad_norm": 0.6671316027641296, |
| "learning_rate": 1.8181818181818183e-07, |
| "loss": 0.056834738701581955, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.05848, |
| "step": 1032, |
| "tokens/total": 4227072, |
| "tokens/train_per_sec_per_gpu": 13.99, |
| "tokens/trainable": 302851 |
| }, |
| { |
| "epoch": 4.919047619047619, |
| "grad_norm": 0.7634382843971252, |
| "learning_rate": 1.722488038277512e-07, |
| "loss": 0.08143440634012222, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.08484, |
| "step": 1033, |
| "tokens/total": 4231168, |
| "tokens/train_per_sec_per_gpu": 14.69, |
| "tokens/trainable": 303195 |
| }, |
| { |
| "epoch": 4.923809523809524, |
| "grad_norm": 0.6088545322418213, |
| "learning_rate": 1.6267942583732058e-07, |
| "loss": 0.04300892725586891, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04395, |
| "step": 1034, |
| "tokens/total": 4235264, |
| "tokens/train_per_sec_per_gpu": 13.65, |
| "tokens/trainable": 303486 |
| }, |
| { |
| "epoch": 4.928571428571429, |
| "grad_norm": 0.6581453084945679, |
| "learning_rate": 1.5311004784688997e-07, |
| "loss": 0.04637235030531883, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04746, |
| "step": 1035, |
| "tokens/total": 4239360, |
| "tokens/train_per_sec_per_gpu": 12.31, |
| "tokens/trainable": 303779 |
| }, |
| { |
| "epoch": 4.933333333333334, |
| "grad_norm": 0.7005976438522339, |
| "learning_rate": 1.4354066985645936e-07, |
| "loss": 0.04131925106048584, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04218, |
| "step": 1036, |
| "tokens/total": 4243456, |
| "tokens/train_per_sec_per_gpu": 12.64, |
| "tokens/trainable": 304030 |
| }, |
| { |
| "epoch": 4.938095238095238, |
| "grad_norm": 0.7025570869445801, |
| "learning_rate": 1.3397129186602872e-07, |
| "loss": 0.07363530248403549, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07641, |
| "step": 1037, |
| "tokens/total": 4247552, |
| "tokens/train_per_sec_per_gpu": 10.85, |
| "tokens/trainable": 304345 |
| }, |
| { |
| "epoch": 4.942857142857143, |
| "grad_norm": 0.7544583082199097, |
| "learning_rate": 1.244019138755981e-07, |
| "loss": 0.07454663515090942, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.0774, |
| "step": 1038, |
| "tokens/total": 4251648, |
| "tokens/train_per_sec_per_gpu": 12.9, |
| "tokens/trainable": 304695 |
| }, |
| { |
| "epoch": 4.947619047619048, |
| "grad_norm": 0.9128358960151672, |
| "learning_rate": 1.1483253588516747e-07, |
| "loss": 0.10159115493297577, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.10693, |
| "step": 1039, |
| "tokens/total": 4255744, |
| "tokens/train_per_sec_per_gpu": 12.82, |
| "tokens/trainable": 304963 |
| }, |
| { |
| "epoch": 4.9523809523809526, |
| "grad_norm": 1.099030613899231, |
| "learning_rate": 1.0526315789473685e-07, |
| "loss": 0.06357499212026596, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06564, |
| "step": 1040, |
| "tokens/total": 4259840, |
| "tokens/train_per_sec_per_gpu": 13.67, |
| "tokens/trainable": 305303 |
| }, |
| { |
| "epoch": 4.957142857142857, |
| "grad_norm": 0.7283017635345459, |
| "learning_rate": 9.569377990430622e-08, |
| "loss": 0.045570287853479385, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.04662, |
| "step": 1041, |
| "tokens/total": 4263936, |
| "tokens/train_per_sec_per_gpu": 13.28, |
| "tokens/trainable": 305575 |
| }, |
| { |
| "epoch": 4.961904761904762, |
| "grad_norm": 0.6875465512275696, |
| "learning_rate": 8.61244019138756e-08, |
| "loss": 0.05585762858390808, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.05745, |
| "step": 1042, |
| "tokens/total": 4268032, |
| "tokens/train_per_sec_per_gpu": 7.66, |
| "tokens/trainable": 305801 |
| }, |
| { |
| "epoch": 4.966666666666667, |
| "grad_norm": 0.6573634743690491, |
| "learning_rate": 7.655502392344498e-08, |
| "loss": 0.039641834795475006, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.04044, |
| "step": 1043, |
| "tokens/total": 4272128, |
| "tokens/train_per_sec_per_gpu": 13.86, |
| "tokens/trainable": 306125 |
| }, |
| { |
| "epoch": 4.9714285714285715, |
| "grad_norm": 1.2348220348358154, |
| "learning_rate": 6.698564593301436e-08, |
| "loss": 0.10564305633306503, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.11143, |
| "step": 1044, |
| "tokens/total": 4276224, |
| "tokens/train_per_sec_per_gpu": 11.01, |
| "tokens/trainable": 306376 |
| }, |
| { |
| "epoch": 4.976190476190476, |
| "grad_norm": 0.8296716213226318, |
| "learning_rate": 5.7416267942583735e-08, |
| "loss": 0.09174297004938126, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.09608, |
| "step": 1045, |
| "tokens/total": 4280320, |
| "tokens/train_per_sec_per_gpu": 13.93, |
| "tokens/trainable": 306707 |
| }, |
| { |
| "epoch": 4.980952380952381, |
| "grad_norm": 0.3389149010181427, |
| "learning_rate": 4.784688995215311e-08, |
| "loss": 0.04257689043879509, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.0435, |
| "step": 1046, |
| "tokens/total": 4284416, |
| "tokens/train_per_sec_per_gpu": 14.2, |
| "tokens/trainable": 307068 |
| }, |
| { |
| "epoch": 4.985714285714286, |
| "grad_norm": 0.8198398947715759, |
| "learning_rate": 3.827751196172249e-08, |
| "loss": 0.060687679797410965, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.06257, |
| "step": 1047, |
| "tokens/total": 4288512, |
| "tokens/train_per_sec_per_gpu": 13.78, |
| "tokens/trainable": 307341 |
| }, |
| { |
| "epoch": 4.9904761904761905, |
| "grad_norm": 0.6749662756919861, |
| "learning_rate": 2.8708133971291867e-08, |
| "loss": 0.06978276371955872, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.55, |
| "memory/max_allocated (GiB)": 38.55, |
| "ppl": 1.07228, |
| "step": 1048, |
| "tokens/total": 4292608, |
| "tokens/train_per_sec_per_gpu": 11.42, |
| "tokens/trainable": 307636 |
| }, |
| { |
| "epoch": 4.995238095238095, |
| "grad_norm": 0.9570760726928711, |
| "learning_rate": 1.9138755980861246e-08, |
| "loss": 0.06977870315313339, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07227, |
| "step": 1049, |
| "tokens/total": 4296704, |
| "tokens/train_per_sec_per_gpu": 11.68, |
| "tokens/trainable": 307883 |
| }, |
| { |
| "epoch": 5.0, |
| "grad_norm": 0.903021514415741, |
| "learning_rate": 9.569377990430623e-09, |
| "loss": 0.06990804523229599, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 38.54, |
| "memory/max_allocated (GiB)": 38.54, |
| "ppl": 1.07241, |
| "step": 1050, |
| "tokens/total": 4300800, |
| "tokens/train_per_sec_per_gpu": 10.28, |
| "tokens/trainable": 308124 |
| }, |
| { |
| "epoch": 5.0, |
| "eval_loss": 0.07747567445039749, |
| "eval_ppl": 1.08056, |
| "eval_runtime": 10.3083, |
| "eval_samples_per_second": 8.634, |
| "eval_steps_per_second": 2.231, |
| "memory/device_reserved (GiB)": 39.21, |
| "memory/max_active (GiB)": 35.58, |
| "memory/max_allocated (GiB)": 35.58, |
| "step": 1050 |
| } |
| ], |
| "logging_steps": 1, |
| "max_steps": 1050, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 5, |
| "save_steps": 210, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 7.76947906904064e+17, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
