| { |
| "best_global_step": 100, |
| "best_metric": 0.9651781320571899, |
| "best_model_checkpoint": "/data6/jongbinwon/evolving_MoE/checkpoints/sft_llama31_8b_sni_domain-c_58325_ffn_only_5ep/checkpoint-100", |
| "epoch": 5.0, |
| "eval_steps": 100, |
| "global_step": 1850, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "entropy": 1.723238617181778, |
| "epoch": 0.027091093802912292, |
| "grad_norm": 7.612579345703125, |
| "learning_rate": 3.2142857142857147e-06, |
| "loss": 2.066279411315918, |
| "mean_token_accuracy": 0.448427795805037, |
| "num_tokens": 21087.0, |
| "step": 10 |
| }, |
| { |
| "entropy": 1.7370752662420272, |
| "epoch": 0.054182187605824585, |
| "grad_norm": 7.184020519256592, |
| "learning_rate": 6.785714285714287e-06, |
| "loss": 2.6899890899658203, |
| "mean_token_accuracy": 0.42496111439540984, |
| "num_tokens": 41335.0, |
| "step": 20 |
| }, |
| { |
| "entropy": 1.843221677839756, |
| "epoch": 0.08127328140873688, |
| "grad_norm": 6.231059551239014, |
| "learning_rate": 1.0357142857142859e-05, |
| "loss": 2.398647117614746, |
| "mean_token_accuracy": 0.5065225204452872, |
| "num_tokens": 61686.0, |
| "step": 30 |
| }, |
| { |
| "entropy": 1.9891791179776193, |
| "epoch": 0.10836437521164917, |
| "grad_norm": 4.975241184234619, |
| "learning_rate": 1.3928571428571429e-05, |
| "loss": 2.2226634979248048, |
| "mean_token_accuracy": 0.6003633574582636, |
| "num_tokens": 81480.0, |
| "step": 40 |
| }, |
| { |
| "entropy": 2.2568055972456933, |
| "epoch": 0.13545546901456146, |
| "grad_norm": 1.7063878774642944, |
| "learning_rate": 1.7500000000000002e-05, |
| "loss": 2.1289892196655273, |
| "mean_token_accuracy": 0.6060962343588472, |
| "num_tokens": 102256.0, |
| "step": 50 |
| }, |
| { |
| "entropy": 2.282596044242382, |
| "epoch": 0.16254656281747376, |
| "grad_norm": 4.551100254058838, |
| "learning_rate": 1.9999862004036567e-05, |
| "loss": 1.528597068786621, |
| "mean_token_accuracy": 0.6501126736402512, |
| "num_tokens": 120836.0, |
| "step": 60 |
| }, |
| { |
| "entropy": 2.3089130371809006, |
| "epoch": 0.18963765662038604, |
| "grad_norm": 4.84531307220459, |
| "learning_rate": 1.9997408848413494e-05, |
| "loss": 1.4672578811645507, |
| "mean_token_accuracy": 0.6635637952014803, |
| "num_tokens": 142008.0, |
| "step": 70 |
| }, |
| { |
| "entropy": 2.4036191523075106, |
| "epoch": 0.21672875042329834, |
| "grad_norm": 4.140300750732422, |
| "learning_rate": 1.9991889981715696e-05, |
| "loss": 1.7146993637084962, |
| "mean_token_accuracy": 0.6447907727211714, |
| "num_tokens": 162908.0, |
| "step": 80 |
| }, |
| { |
| "entropy": 2.402925634384155, |
| "epoch": 0.24381984422621064, |
| "grad_norm": 3.240445613861084, |
| "learning_rate": 1.9983307096306773e-05, |
| "loss": 1.5633929252624512, |
| "mean_token_accuracy": 0.6392208762466908, |
| "num_tokens": 183267.0, |
| "step": 90 |
| }, |
| { |
| "entropy": 2.254238374531269, |
| "epoch": 0.2709109380291229, |
| "grad_norm": 2.4666740894317627, |
| "learning_rate": 1.997166282413338e-05, |
| "loss": 1.3440716743469239, |
| "mean_token_accuracy": 0.66012231297791, |
| "num_tokens": 203918.0, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.2709109380291229, |
| "eval_entropy": 2.094589021988213, |
| "eval_loss": 0.9651781320571899, |
| "eval_mean_token_accuracy": 0.764254976878874, |
| "eval_num_tokens": 203918.0, |
| "eval_runtime": 25.5717, |
| "eval_samples_per_second": 20.022, |
| "eval_steps_per_second": 10.011, |
| "step": 100 |
| }, |
| { |
| "entropy": 2.216274145245552, |
| "epoch": 0.2980020318320352, |
| "grad_norm": 3.9435603618621826, |
| "learning_rate": 1.995696073591817e-05, |
| "loss": 1.493989849090576, |
| "mean_token_accuracy": 0.6519980503246188, |
| "num_tokens": 222513.0, |
| "step": 110 |
| }, |
| { |
| "entropy": 2.2828305318951605, |
| "epoch": 0.3250931256349475, |
| "grad_norm": 3.6630003452301025, |
| "learning_rate": 1.9939205340064793e-05, |
| "loss": 1.638005256652832, |
| "mean_token_accuracy": 0.6362768037244677, |
| "num_tokens": 240852.0, |
| "step": 120 |
| }, |
| { |
| "entropy": 2.2786295488476753, |
| "epoch": 0.3521842194378598, |
| "grad_norm": 4.0613274574279785, |
| "learning_rate": 1.991840208127543e-05, |
| "loss": 1.3484866142272949, |
| "mean_token_accuracy": 0.708053832501173, |
| "num_tokens": 260786.0, |
| "step": 130 |
| }, |
| { |
| "entropy": 2.239742587506771, |
| "epoch": 0.37927531324077207, |
| "grad_norm": 3.4148905277252197, |
| "learning_rate": 1.989455733888115e-05, |
| "loss": 1.4180074691772462, |
| "mean_token_accuracy": 0.6825597375631333, |
| "num_tokens": 280647.0, |
| "step": 140 |
| }, |
| { |
| "entropy": 2.224020117521286, |
| "epoch": 0.4063664070436844, |
| "grad_norm": 4.721391677856445, |
| "learning_rate": 1.986767842488569e-05, |
| "loss": 1.2252963066101075, |
| "mean_token_accuracy": 0.6958539757877589, |
| "num_tokens": 301192.0, |
| "step": 150 |
| }, |
| { |
| "entropy": 2.1590186595916747, |
| "epoch": 0.4334575008465967, |
| "grad_norm": 3.9556772708892822, |
| "learning_rate": 1.9837773581723215e-05, |
| "loss": 1.5099032402038575, |
| "mean_token_accuracy": 0.6419584900140762, |
| "num_tokens": 321679.0, |
| "step": 160 |
| }, |
| { |
| "entropy": 2.2231348380446434, |
| "epoch": 0.460548594649509, |
| "grad_norm": 3.490177869796753, |
| "learning_rate": 1.980485197973079e-05, |
| "loss": 1.6476173400878906, |
| "mean_token_accuracy": 0.6805331652984024, |
| "num_tokens": 342221.0, |
| "step": 170 |
| }, |
| { |
| "entropy": 2.24891133159399, |
| "epoch": 0.4876396884524213, |
| "grad_norm": 4.099713325500488, |
| "learning_rate": 1.9768923714336274e-05, |
| "loss": 1.315092182159424, |
| "mean_token_accuracy": 0.6847983291372657, |
| "num_tokens": 362847.0, |
| "step": 180 |
| }, |
| { |
| "entropy": 2.1428863570094108, |
| "epoch": 0.5147307822553335, |
| "grad_norm": 3.966991901397705, |
| "learning_rate": 1.9729999802962555e-05, |
| "loss": 1.2681650161743163, |
| "mean_token_accuracy": 0.7244645357131958, |
| "num_tokens": 384086.0, |
| "step": 190 |
| }, |
| { |
| "entropy": 2.2040064856410027, |
| "epoch": 0.5418218760582458, |
| "grad_norm": 3.2430970668792725, |
| "learning_rate": 1.9688092181649065e-05, |
| "loss": 1.3249666213989257, |
| "mean_token_accuracy": 0.722496079467237, |
| "num_tokens": 404677.0, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.5418218760582458, |
| "eval_entropy": 2.0920665017329156, |
| "eval_loss": 0.9940700531005859, |
| "eval_mean_token_accuracy": 0.757662381103728, |
| "eval_num_tokens": 404677.0, |
| "eval_runtime": 25.5435, |
| "eval_samples_per_second": 20.044, |
| "eval_steps_per_second": 10.022, |
| "step": 200 |
| }, |
| { |
| "entropy": 2.156522032618523, |
| "epoch": 0.5689129698611581, |
| "grad_norm": 4.980217933654785, |
| "learning_rate": 1.964321370139157e-05, |
| "loss": 1.298858070373535, |
| "mean_token_accuracy": 0.7201794616878032, |
| "num_tokens": 426644.0, |
| "step": 210 |
| }, |
| { |
| "entropy": 2.1643049642443657, |
| "epoch": 0.5960040636640704, |
| "grad_norm": 5.744068145751953, |
| "learning_rate": 1.9595378124201402e-05, |
| "loss": 1.2987921714782715, |
| "mean_token_accuracy": 0.7069474339485169, |
| "num_tokens": 446791.0, |
| "step": 220 |
| }, |
| { |
| "entropy": 2.177510863542557, |
| "epoch": 0.6230951574669827, |
| "grad_norm": 8.363221168518066, |
| "learning_rate": 1.9544600118885356e-05, |
| "loss": 1.4281065940856934, |
| "mean_token_accuracy": 0.7097761157900095, |
| "num_tokens": 466287.0, |
| "step": 230 |
| }, |
| { |
| "entropy": 2.2098993390798567, |
| "epoch": 0.650186251269895, |
| "grad_norm": 4.0296220779418945, |
| "learning_rate": 1.9490895256547465e-05, |
| "loss": 1.0510370254516601, |
| "mean_token_accuracy": 0.7287774980068207, |
| "num_tokens": 485912.0, |
| "step": 240 |
| }, |
| { |
| "entropy": 2.194219082593918, |
| "epoch": 0.6772773450728073, |
| "grad_norm": 4.584537982940674, |
| "learning_rate": 1.9434280005814115e-05, |
| "loss": 1.5170968055725098, |
| "mean_token_accuracy": 0.6617882775142789, |
| "num_tokens": 506698.0, |
| "step": 250 |
| }, |
| { |
| "entropy": 2.1990948766469955, |
| "epoch": 0.7043684388757196, |
| "grad_norm": 2.1026809215545654, |
| "learning_rate": 1.9374771727783956e-05, |
| "loss": 1.2185900688171387, |
| "mean_token_accuracy": 0.6906542837619781, |
| "num_tokens": 526063.0, |
| "step": 260 |
| }, |
| { |
| "entropy": 2.1788215398788453, |
| "epoch": 0.731459532678632, |
| "grad_norm": 3.8161611557006836, |
| "learning_rate": 1.931238867070408e-05, |
| "loss": 1.3875349044799805, |
| "mean_token_accuracy": 0.7175548300147057, |
| "num_tokens": 547940.0, |
| "step": 270 |
| }, |
| { |
| "entropy": 2.1797836825251578, |
| "epoch": 0.7585506264815441, |
| "grad_norm": 4.361836910247803, |
| "learning_rate": 1.924714996437421e-05, |
| "loss": 1.4559088706970216, |
| "mean_token_accuracy": 0.6750684441998601, |
| "num_tokens": 567784.0, |
| "step": 280 |
| }, |
| { |
| "entropy": 2.192278742790222, |
| "epoch": 0.7856417202844564, |
| "grad_norm": 5.223799705505371, |
| "learning_rate": 1.9179075614280513e-05, |
| "loss": 1.3970552444458009, |
| "mean_token_accuracy": 0.7280023746192456, |
| "num_tokens": 587559.0, |
| "step": 290 |
| }, |
| { |
| "entropy": 2.152565708756447, |
| "epoch": 0.8127328140873687, |
| "grad_norm": 4.498564720153809, |
| "learning_rate": 1.9108186495460934e-05, |
| "loss": 1.296685028076172, |
| "mean_token_accuracy": 0.6948160078376532, |
| "num_tokens": 608955.0, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.8127328140873687, |
| "eval_entropy": 2.086765900719911, |
| "eval_loss": 0.9978444576263428, |
| "eval_mean_token_accuracy": 0.7629395253607072, |
| "eval_num_tokens": 608955.0, |
| "eval_runtime": 25.5161, |
| "eval_samples_per_second": 20.066, |
| "eval_steps_per_second": 10.033, |
| "step": 300 |
| }, |
| { |
| "entropy": 2.2242535188794137, |
| "epoch": 0.839823907890281, |
| "grad_norm": 4.154886245727539, |
| "learning_rate": 1.9034504346103825e-05, |
| "loss": 1.5084619522094727, |
| "mean_token_accuracy": 0.699345251917839, |
| "num_tokens": 629498.0, |
| "step": 310 |
| }, |
| { |
| "entropy": 2.2500312089920045, |
| "epoch": 0.8669150016931934, |
| "grad_norm": 3.7556686401367188, |
| "learning_rate": 1.8958051760881938e-05, |
| "loss": 1.3967250823974608, |
| "mean_token_accuracy": 0.6777329007163644, |
| "num_tokens": 648909.0, |
| "step": 320 |
| }, |
| { |
| "entropy": 2.270790347456932, |
| "epoch": 0.8940060954961057, |
| "grad_norm": 5.752708911895752, |
| "learning_rate": 1.8878852184023754e-05, |
| "loss": 1.2833734512329102, |
| "mean_token_accuracy": 0.6765352768823505, |
| "num_tokens": 667376.0, |
| "step": 330 |
| }, |
| { |
| "entropy": 2.1318136006593704, |
| "epoch": 0.921097189299018, |
| "grad_norm": 4.369615077972412, |
| "learning_rate": 1.879692990212428e-05, |
| "loss": 1.2982247352600098, |
| "mean_token_accuracy": 0.7231215512380004, |
| "num_tokens": 689776.0, |
| "step": 340 |
| }, |
| { |
| "entropy": 2.179507865011692, |
| "epoch": 0.9481882831019303, |
| "grad_norm": 3.208723783493042, |
| "learning_rate": 1.8712310036697575e-05, |
| "loss": 1.279671859741211, |
| "mean_token_accuracy": 0.7093656450510025, |
| "num_tokens": 710374.0, |
| "step": 350 |
| }, |
| { |
| "entropy": 2.1694433763623238, |
| "epoch": 0.9752793769048426, |
| "grad_norm": 6.71851110458374, |
| "learning_rate": 1.8625018536473204e-05, |
| "loss": 1.485358428955078, |
| "mean_token_accuracy": 0.6655262919142843, |
| "num_tokens": 731208.0, |
| "step": 360 |
| }, |
| { |
| "entropy": 2.1039708019935923, |
| "epoch": 1.0, |
| "grad_norm": 5.395904064178467, |
| "learning_rate": 1.8535082169439057e-05, |
| "loss": 1.3225526809692383, |
| "mean_token_accuracy": 0.7076871527792656, |
| "num_tokens": 750475.0, |
| "step": 370 |
| }, |
| { |
| "entropy": 2.1732485339045526, |
| "epoch": 1.0270910938029123, |
| "grad_norm": 3.098240852355957, |
| "learning_rate": 1.844252851463292e-05, |
| "loss": 1.514673614501953, |
| "mean_token_accuracy": 0.7053313782438636, |
| "num_tokens": 772399.0, |
| "step": 380 |
| }, |
| { |
| "entropy": 2.1763458222150804, |
| "epoch": 1.0541821876058246, |
| "grad_norm": 3.707331895828247, |
| "learning_rate": 1.834738595368536e-05, |
| "loss": 1.0740927696228026, |
| "mean_token_accuracy": 0.7563056144863367, |
| "num_tokens": 792378.0, |
| "step": 390 |
| }, |
| { |
| "entropy": 2.171446980535984, |
| "epoch": 1.081273281408737, |
| "grad_norm": 5.769153594970703, |
| "learning_rate": 1.824968366211648e-05, |
| "loss": 1.2906767845153808, |
| "mean_token_accuracy": 0.7230573301203549, |
| "num_tokens": 811796.0, |
| "step": 400 |
| }, |
| { |
| "epoch": 1.081273281408737, |
| "eval_entropy": 2.0477666831575334, |
| "eval_loss": 1.0081422328948975, |
| "eval_mean_token_accuracy": 0.7632476337021217, |
| "eval_num_tokens": 811796.0, |
| "eval_runtime": 25.4904, |
| "eval_samples_per_second": 20.086, |
| "eval_steps_per_second": 10.043, |
| "step": 400 |
| }, |
| { |
| "entropy": 2.1150058075785636, |
| "epoch": 1.1083643752116492, |
| "grad_norm": 4.51067590713501, |
| "learning_rate": 1.8149451600389207e-05, |
| "loss": 1.3508639335632324, |
| "mean_token_accuracy": 0.7513733027502895, |
| "num_tokens": 832769.0, |
| "step": 410 |
| }, |
| { |
| "entropy": 2.1037127494812013, |
| "epoch": 1.1354554690145615, |
| "grad_norm": 2.397596836090088, |
| "learning_rate": 1.8046720504721938e-05, |
| "loss": 1.0258020401000976, |
| "mean_token_accuracy": 0.74247965849936, |
| "num_tokens": 853670.0, |
| "step": 420 |
| }, |
| { |
| "entropy": 2.0551841765642167, |
| "epoch": 1.1625465628174738, |
| "grad_norm": 9.206628799438477, |
| "learning_rate": 1.794152187766323e-05, |
| "loss": 1.1884271621704101, |
| "mean_token_accuracy": 0.7030579019337893, |
| "num_tokens": 874009.0, |
| "step": 430 |
| }, |
| { |
| "entropy": 2.0734852880239485, |
| "epoch": 1.1896376566203861, |
| "grad_norm": 4.471444129943848, |
| "learning_rate": 1.783388797843154e-05, |
| "loss": 1.2394503593444823, |
| "mean_token_accuracy": 0.7340864278376102, |
| "num_tokens": 895613.0, |
| "step": 440 |
| }, |
| { |
| "entropy": 2.134996885061264, |
| "epoch": 1.2167287504232984, |
| "grad_norm": 6.01848840713501, |
| "learning_rate": 1.77238518130229e-05, |
| "loss": 1.227487850189209, |
| "mean_token_accuracy": 0.7364656325429678, |
| "num_tokens": 914468.0, |
| "step": 450 |
| }, |
| { |
| "entropy": 2.061158812046051, |
| "epoch": 1.2438198442262107, |
| "grad_norm": 12.69832992553711, |
| "learning_rate": 1.761144712408965e-05, |
| "loss": 1.221731185913086, |
| "mean_token_accuracy": 0.7208205627277493, |
| "num_tokens": 936076.0, |
| "step": 460 |
| }, |
| { |
| "entropy": 2.1015277206897736, |
| "epoch": 1.270910938029123, |
| "grad_norm": 3.7558536529541016, |
| "learning_rate": 1.749670838059318e-05, |
| "loss": 0.9690493583679199, |
| "mean_token_accuracy": 0.7604702278971672, |
| "num_tokens": 956253.0, |
| "step": 470 |
| }, |
| { |
| "entropy": 2.1128511682152746, |
| "epoch": 1.298002031832035, |
| "grad_norm": 7.6537275314331055, |
| "learning_rate": 1.7379670767234045e-05, |
| "loss": 1.0514067649841308, |
| "mean_token_accuracy": 0.7440206557512283, |
| "num_tokens": 976137.0, |
| "step": 480 |
| }, |
| { |
| "entropy": 2.106458379328251, |
| "epoch": 1.3250931256349476, |
| "grad_norm": 7.5504045486450195, |
| "learning_rate": 1.7260370173662543e-05, |
| "loss": 1.0261421203613281, |
| "mean_token_accuracy": 0.7460719928145408, |
| "num_tokens": 995767.0, |
| "step": 490 |
| }, |
| { |
| "entropy": 2.118429370224476, |
| "epoch": 1.3521842194378597, |
| "grad_norm": 4.931035995483398, |
| "learning_rate": 1.7138843183473127e-05, |
| "loss": 1.5636432647705079, |
| "mean_token_accuracy": 0.7403618179261684, |
| "num_tokens": 1015342.0, |
| "step": 500 |
| }, |
| { |
| "epoch": 1.3521842194378597, |
| "eval_entropy": 1.9999370207078755, |
| "eval_loss": 1.032294511795044, |
| "eval_mean_token_accuracy": 0.7589045927161351, |
| "eval_num_tokens": 1015342.0, |
| "eval_runtime": 25.5016, |
| "eval_samples_per_second": 20.077, |
| "eval_steps_per_second": 10.039, |
| "step": 500 |
| }, |
| { |
| "entropy": 2.124773620069027, |
| "epoch": 1.379275313240772, |
| "grad_norm": 10.168334007263184, |
| "learning_rate": 1.7015127062986044e-05, |
| "loss": 0.8846927642822265, |
| "mean_token_accuracy": 0.7411885142326355, |
| "num_tokens": 1034010.0, |
| "step": 510 |
| }, |
| { |
| "entropy": 2.0986449405550958, |
| "epoch": 1.4063664070436843, |
| "grad_norm": 6.97003698348999, |
| "learning_rate": 1.6889259749819568e-05, |
| "loss": 1.023176383972168, |
| "mean_token_accuracy": 0.7303699310868979, |
| "num_tokens": 1053408.0, |
| "step": 520 |
| }, |
| { |
| "entropy": 2.1111062809824945, |
| "epoch": 1.4334575008465966, |
| "grad_norm": 5.16034460067749, |
| "learning_rate": 1.6761279841256437e-05, |
| "loss": 1.3746380805969238, |
| "mean_token_accuracy": 0.7154265632852912, |
| "num_tokens": 1074316.0, |
| "step": 530 |
| }, |
| { |
| "entropy": 2.071941052377224, |
| "epoch": 1.460548594649509, |
| "grad_norm": 5.507922649383545, |
| "learning_rate": 1.6631226582407954e-05, |
| "loss": 1.3437743186950684, |
| "mean_token_accuracy": 0.6885566545650363, |
| "num_tokens": 1094208.0, |
| "step": 540 |
| }, |
| { |
| "entropy": 2.1154344886541367, |
| "epoch": 1.4876396884524212, |
| "grad_norm": 3.5610263347625732, |
| "learning_rate": 1.649913985417946e-05, |
| "loss": 1.1012388229370118, |
| "mean_token_accuracy": 0.7634946122765541, |
| "num_tokens": 1113777.0, |
| "step": 550 |
| }, |
| { |
| "entropy": 2.0671692818403242, |
| "epoch": 1.5147307822553335, |
| "grad_norm": 8.337108612060547, |
| "learning_rate": 1.636506016104084e-05, |
| "loss": 1.3132530212402345, |
| "mean_token_accuracy": 0.7397821169346571, |
| "num_tokens": 1133996.0, |
| "step": 560 |
| }, |
| { |
| "entropy": 2.0944964781403543, |
| "epoch": 1.5418218760582458, |
| "grad_norm": 7.714909076690674, |
| "learning_rate": 1.6229028618605776e-05, |
| "loss": 0.8892477035522461, |
| "mean_token_accuracy": 0.7313346045091749, |
| "num_tokens": 1154047.0, |
| "step": 570 |
| }, |
| { |
| "entropy": 2.0559984400868414, |
| "epoch": 1.5689129698611581, |
| "grad_norm": 6.229889392852783, |
| "learning_rate": 1.609108694102366e-05, |
| "loss": 0.9835987091064453, |
| "mean_token_accuracy": 0.7751555263996124, |
| "num_tokens": 1174616.0, |
| "step": 580 |
| }, |
| { |
| "entropy": 2.054904416203499, |
| "epoch": 1.5960040636640704, |
| "grad_norm": 3.5100467205047607, |
| "learning_rate": 1.59512774281879e-05, |
| "loss": 0.7898230075836181, |
| "mean_token_accuracy": 0.7902765288949013, |
| "num_tokens": 1194173.0, |
| "step": 590 |
| }, |
| { |
| "entropy": 2.0093825444579125, |
| "epoch": 1.6230951574669827, |
| "grad_norm": 5.0431742668151855, |
| "learning_rate": 1.580964295276463e-05, |
| "loss": 1.0017314910888673, |
| "mean_token_accuracy": 0.7571999192237854, |
| "num_tokens": 1214366.0, |
| "step": 600 |
| }, |
| { |
| "epoch": 1.6230951574669827, |
| "eval_entropy": 1.9694703868590295, |
| "eval_loss": 1.0424904823303223, |
| "eval_mean_token_accuracy": 0.7560218448052183, |
| "eval_num_tokens": 1214366.0, |
| "eval_runtime": 25.5499, |
| "eval_samples_per_second": 20.039, |
| "eval_steps_per_second": 10.02, |
| "step": 600 |
| }, |
| { |
| "entropy": 2.0107428073883056, |
| "epoch": 1.650186251269895, |
| "grad_norm": 7.987220287322998, |
| "learning_rate": 1.5666226947045814e-05, |
| "loss": 1.3089231491088866, |
| "mean_token_accuracy": 0.7401313630864024, |
| "num_tokens": 1236562.0, |
| "step": 610 |
| }, |
| { |
| "entropy": 2.046839062869549, |
| "epoch": 1.6772773450728073, |
| "grad_norm": 6.632065296173096, |
| "learning_rate": 1.552107338963067e-05, |
| "loss": 0.9845743179321289, |
| "mean_token_accuracy": 0.755998631566763, |
| "num_tokens": 1257112.0, |
| "step": 620 |
| }, |
| { |
| "entropy": 2.1148023292422296, |
| "epoch": 1.7043684388757196, |
| "grad_norm": 8.282198905944824, |
| "learning_rate": 1.5374226791939628e-05, |
| "loss": 1.3935293197631835, |
| "mean_token_accuracy": 0.7302353219129145, |
| "num_tokens": 1276577.0, |
| "step": 630 |
| }, |
| { |
| "entropy": 2.0933067336678506, |
| "epoch": 1.731459532678632, |
| "grad_norm": 5.363010406494141, |
| "learning_rate": 1.5225732184564886e-05, |
| "loss": 1.2860239028930665, |
| "mean_token_accuracy": 0.7248616352677345, |
| "num_tokens": 1296086.0, |
| "step": 640 |
| }, |
| { |
| "entropy": 2.1054985627532004, |
| "epoch": 1.758550626481544, |
| "grad_norm": 4.800771236419678, |
| "learning_rate": 1.5075635103461731e-05, |
| "loss": 1.2266152381896973, |
| "mean_token_accuracy": 0.7662290595471859, |
| "num_tokens": 1316442.0, |
| "step": 650 |
| }, |
| { |
| "entropy": 2.1522702112793923, |
| "epoch": 1.7856417202844566, |
| "grad_norm": 3.5823662281036377, |
| "learning_rate": 1.4923981575984936e-05, |
| "loss": 1.0100022315979005, |
| "mean_token_accuracy": 0.7728497218340635, |
| "num_tokens": 1337169.0, |
| "step": 660 |
| }, |
| { |
| "entropy": 2.1094584688544273, |
| "epoch": 1.8127328140873686, |
| "grad_norm": 5.589018821716309, |
| "learning_rate": 1.4770818106774427e-05, |
| "loss": 1.0324097633361817, |
| "mean_token_accuracy": 0.7925379849970341, |
| "num_tokens": 1357759.0, |
| "step": 670 |
| }, |
| { |
| "entropy": 2.182632897794247, |
| "epoch": 1.8398239078902812, |
| "grad_norm": 7.491511344909668, |
| "learning_rate": 1.4616191663494619e-05, |
| "loss": 1.3248936653137207, |
| "mean_token_accuracy": 0.7274331038817763, |
| "num_tokens": 1378191.0, |
| "step": 680 |
| }, |
| { |
| "entropy": 2.1101855129003524, |
| "epoch": 1.8669150016931932, |
| "grad_norm": 6.318648815155029, |
| "learning_rate": 1.4460149662431749e-05, |
| "loss": 1.054247760772705, |
| "mean_token_accuracy": 0.7554519359022379, |
| "num_tokens": 1397881.0, |
| "step": 690 |
| }, |
| { |
| "entropy": 2.096898023784161, |
| "epoch": 1.8940060954961058, |
| "grad_norm": 6.2741875648498535, |
| "learning_rate": 1.430273995395365e-05, |
| "loss": 1.0373686790466308, |
| "mean_token_accuracy": 0.7596361611038447, |
| "num_tokens": 1418877.0, |
| "step": 700 |
| }, |
| { |
| "epoch": 1.8940060954961058, |
| "eval_entropy": 2.013828214723617, |
| "eval_loss": 1.0287954807281494, |
| "eval_mean_token_accuracy": 0.7496589213842526, |
| "eval_num_tokens": 1418877.0, |
| "eval_runtime": 25.4995, |
| "eval_samples_per_second": 20.079, |
| "eval_steps_per_second": 10.039, |
| "step": 700 |
| }, |
| { |
| "entropy": 2.0728593289852144, |
| "epoch": 1.9210971892990178, |
| "grad_norm": 6.076820373535156, |
| "learning_rate": 1.4144010807836412e-05, |
| "loss": 1.0167503356933594, |
| "mean_token_accuracy": 0.7612185761332512, |
| "num_tokens": 1439198.0, |
| "step": 710 |
| }, |
| { |
| "entropy": 2.11173554956913, |
| "epoch": 1.9481882831019304, |
| "grad_norm": 6.451402187347412, |
| "learning_rate": 1.3984010898462417e-05, |
| "loss": 1.089035415649414, |
| "mean_token_accuracy": 0.7369910277426243, |
| "num_tokens": 1461226.0, |
| "step": 720 |
| }, |
| { |
| "entropy": 2.0723678693175316, |
| "epoch": 1.9752793769048425, |
| "grad_norm": 5.50106954574585, |
| "learning_rate": 1.3822789289894316e-05, |
| "loss": 0.9883082389831543, |
| "mean_token_accuracy": 0.7898955499753356, |
| "num_tokens": 1482366.0, |
| "step": 730 |
| }, |
| { |
| "entropy": 2.0618706647663902, |
| "epoch": 2.0, |
| "grad_norm": 50.42632293701172, |
| "learning_rate": 1.3660395420829516e-05, |
| "loss": 0.8657957077026367, |
| "mean_token_accuracy": 0.7464919359716651, |
| "num_tokens": 1500950.0, |
| "step": 740 |
| }, |
| { |
| "entropy": 2.050840379297733, |
| "epoch": 2.027091093802912, |
| "grad_norm": 6.361630439758301, |
| "learning_rate": 1.3496879089439761e-05, |
| "loss": 0.8690568923950195, |
| "mean_token_accuracy": 0.7947060691192747, |
| "num_tokens": 1519798.0, |
| "step": 750 |
| }, |
| { |
| "entropy": 1.926042178273201, |
| "epoch": 2.0541821876058246, |
| "grad_norm": 2.3748672008514404, |
| "learning_rate": 1.3332290438100508e-05, |
| "loss": 0.8486003875732422, |
| "mean_token_accuracy": 0.8309563081711531, |
| "num_tokens": 1539927.0, |
| "step": 760 |
| }, |
| { |
| "entropy": 1.8769470036029816, |
| "epoch": 2.0812732814087367, |
| "grad_norm": 8.954164505004883, |
| "learning_rate": 1.3166679938014728e-05, |
| "loss": 0.7576936244964599, |
| "mean_token_accuracy": 0.8146512515842914, |
| "num_tokens": 1560176.0, |
| "step": 770 |
| }, |
| { |
| "entropy": 1.8487665757536889, |
| "epoch": 2.108364375211649, |
| "grad_norm": 8.518715858459473, |
| "learning_rate": 1.3000098373735885e-05, |
| "loss": 0.9359015464782715, |
| "mean_token_accuracy": 0.788595624268055, |
| "num_tokens": 1579888.0, |
| "step": 780 |
| }, |
| { |
| "entropy": 1.8781982406973838, |
| "epoch": 2.1354554690145613, |
| "grad_norm": 5.680595397949219, |
| "learning_rate": 1.283259682759484e-05, |
| "loss": 0.9370273590087891, |
| "mean_token_accuracy": 0.7608782526105642, |
| "num_tokens": 1601141.0, |
| "step": 790 |
| }, |
| { |
| "entropy": 1.9456540763378143, |
| "epoch": 2.162546562817474, |
| "grad_norm": 5.94552755355835, |
| "learning_rate": 1.266422666403539e-05, |
| "loss": 0.9432417869567871, |
| "mean_token_accuracy": 0.8288755964487791, |
| "num_tokens": 1621542.0, |
| "step": 800 |
| }, |
| { |
| "epoch": 2.162546562817474, |
| "eval_entropy": 1.9436165052466094, |
| "eval_loss": 1.06016206741333, |
| "eval_mean_token_accuracy": 0.7518258688505739, |
| "eval_num_tokens": 1621542.0, |
| "eval_runtime": 25.5171, |
| "eval_samples_per_second": 20.065, |
| "eval_steps_per_second": 10.032, |
| "step": 800 |
| }, |
| { |
| "entropy": 1.9728469848632812, |
| "epoch": 2.189637656620386, |
| "grad_norm": 7.825141906738281, |
| "learning_rate": 1.2495039513863378e-05, |
| "loss": 1.036830997467041, |
| "mean_token_accuracy": 0.8224836468696595, |
| "num_tokens": 1641245.0, |
| "step": 810 |
| }, |
| { |
| "entropy": 1.9231234237551689, |
| "epoch": 2.2167287504232984, |
| "grad_norm": 6.70737361907959, |
| "learning_rate": 1.2325087258414039e-05, |
| "loss": 0.7759229183197022, |
| "mean_token_accuracy": 0.8049912886694074, |
| "num_tokens": 1663067.0, |
| "step": 820 |
| }, |
| { |
| "entropy": 1.8155759632587434, |
| "epoch": 2.2438198442262105, |
| "grad_norm": 9.183692932128906, |
| "learning_rate": 1.2154422013642575e-05, |
| "loss": 0.8489904403686523, |
| "mean_token_accuracy": 0.8039800453931093, |
| "num_tokens": 1683171.0, |
| "step": 830 |
| }, |
| { |
| "entropy": 1.9373584493994713, |
| "epoch": 2.270910938029123, |
| "grad_norm": 6.903590679168701, |
| "learning_rate": 1.1983096114142773e-05, |
| "loss": 1.001230525970459, |
| "mean_token_accuracy": 0.7758175190538168, |
| "num_tokens": 1702798.0, |
| "step": 840 |
| }, |
| { |
| "entropy": 2.008491799235344, |
| "epoch": 2.298002031832035, |
| "grad_norm": 7.595852375030518, |
| "learning_rate": 1.1811162097098559e-05, |
| "loss": 0.9919692039489746, |
| "mean_token_accuracy": 0.8151415549218655, |
| "num_tokens": 1722315.0, |
| "step": 850 |
| }, |
| { |
| "entropy": 1.9535631239414215, |
| "epoch": 2.3250931256349476, |
| "grad_norm": 8.800715446472168, |
| "learning_rate": 1.1638672686173453e-05, |
| "loss": 0.7722007274627686, |
| "mean_token_accuracy": 0.8053982924669981, |
| "num_tokens": 1741922.0, |
| "step": 860 |
| }, |
| { |
| "entropy": 1.9666619777679444, |
| "epoch": 2.3521842194378597, |
| "grad_norm": 5.499095916748047, |
| "learning_rate": 1.1465680775342821e-05, |
| "loss": 1.0401187896728517, |
| "mean_token_accuracy": 0.7628452027216553, |
| "num_tokens": 1761561.0, |
| "step": 870 |
| }, |
| { |
| "entropy": 1.9506760001182557, |
| "epoch": 2.3792753132407722, |
| "grad_norm": 15.519598007202148, |
| "learning_rate": 1.1292239412673906e-05, |
| "loss": 0.9919161796569824, |
| "mean_token_accuracy": 0.7912353215739131, |
| "num_tokens": 1781403.0, |
| "step": 880 |
| }, |
| { |
| "entropy": 1.8824239298701286, |
| "epoch": 2.4063664070436843, |
| "grad_norm": 8.007031440734863, |
| "learning_rate": 1.1118401784058612e-05, |
| "loss": 0.8426601409912109, |
| "mean_token_accuracy": 0.8043077282607556, |
| "num_tokens": 1803471.0, |
| "step": 890 |
| }, |
| { |
| "entropy": 1.9203020244836808, |
| "epoch": 2.433457500846597, |
| "grad_norm": 8.088946342468262, |
| "learning_rate": 1.0944221196904029e-05, |
| "loss": 1.0317404747009278, |
| "mean_token_accuracy": 0.7911908756941557, |
| "num_tokens": 1823668.0, |
| "step": 900 |
| }, |
| { |
| "epoch": 2.433457500846597, |
| "eval_entropy": 1.8476642882451415, |
| "eval_loss": 1.102004051208496, |
| "eval_mean_token_accuracy": 0.7586654123151675, |
| "eval_num_tokens": 1823668.0, |
| "eval_runtime": 25.5319, |
| "eval_samples_per_second": 20.053, |
| "eval_steps_per_second": 10.027, |
| "step": 900 |
| }, |
| { |
| "entropy": 1.8842667803168296, |
| "epoch": 2.460548594649509, |
| "grad_norm": 10.67811393737793, |
| "learning_rate": 1.0769751063785661e-05, |
| "loss": 0.7327501773834229, |
| "mean_token_accuracy": 0.7824661247432232, |
| "num_tokens": 1844175.0, |
| "step": 910 |
| }, |
| { |
| "entropy": 1.8312113255262374, |
| "epoch": 2.4876396884524214, |
| "grad_norm": 5.517256736755371, |
| "learning_rate": 1.0595044886068457e-05, |
| "loss": 0.6679872512817383, |
| "mean_token_accuracy": 0.8325993716716766, |
| "num_tokens": 1864323.0, |
| "step": 920 |
| }, |
| { |
| "entropy": 1.820173794031143, |
| "epoch": 2.5147307822553335, |
| "grad_norm": 5.745193958282471, |
| "learning_rate": 1.0420156237500571e-05, |
| "loss": 0.6937276363372803, |
| "mean_token_accuracy": 0.8053716959431767, |
| "num_tokens": 1884326.0, |
| "step": 930 |
| }, |
| { |
| "entropy": 1.8793843016028404, |
| "epoch": 2.541821876058246, |
| "grad_norm": 7.830550670623779, |
| "learning_rate": 1.0245138747784953e-05, |
| "loss": 0.8545784950256348, |
| "mean_token_accuracy": 0.8472533781081438, |
| "num_tokens": 1904095.0, |
| "step": 940 |
| }, |
| { |
| "entropy": 1.8013970047235488, |
| "epoch": 2.568912969861158, |
| "grad_norm": 6.052741527557373, |
| "learning_rate": 1.007004608613375e-05, |
| "loss": 0.7021704196929932, |
| "mean_token_accuracy": 0.8472875323146581, |
| "num_tokens": 1924391.0, |
| "step": 950 |
| }, |
| { |
| "entropy": 1.9338349267840385, |
| "epoch": 2.59600406366407, |
| "grad_norm": 12.163394927978516, |
| "learning_rate": 9.89493194481064e-06, |
| "loss": 0.9507933616638183, |
| "mean_token_accuracy": 0.807798033207655, |
| "num_tokens": 1944746.0, |
| "step": 960 |
| }, |
| { |
| "entropy": 1.879988330602646, |
| "epoch": 2.6230951574669827, |
| "grad_norm": 6.659127712249756, |
| "learning_rate": 9.719850022666044e-06, |
| "loss": 0.6338046550750732, |
| "mean_token_accuracy": 0.8415262099355459, |
| "num_tokens": 1967427.0, |
| "step": 970 |
| }, |
| { |
| "entropy": 1.890741939842701, |
| "epoch": 2.6501862512698953, |
| "grad_norm": 9.862802505493164, |
| "learning_rate": 9.544854008670366e-06, |
| "loss": 0.7429360389709473, |
| "mean_token_accuracy": 0.8043466622009874, |
| "num_tokens": 1986505.0, |
| "step": 980 |
| }, |
| { |
| "entropy": 1.8231979936361313, |
| "epoch": 2.6772773450728073, |
| "grad_norm": 11.50146198272705, |
| "learning_rate": 9.369997565450214e-06, |
| "loss": 1.0342053413391112, |
| "mean_token_accuracy": 0.7860685650259256, |
| "num_tokens": 2006357.0, |
| "step": 990 |
| }, |
| { |
| "entropy": 1.8516354262828827, |
| "epoch": 2.7043684388757194, |
| "grad_norm": 3.143589973449707, |
| "learning_rate": 9.195334312832742e-06, |
| "loss": 0.8172811508178711, |
| "mean_token_accuracy": 0.8135180365294218, |
| "num_tokens": 2026498.0, |
| "step": 1000 |
| }, |
| { |
| "epoch": 2.7043684388757194, |
| "eval_entropy": 1.853846381418407, |
| "eval_loss": 1.0932104587554932, |
| "eval_mean_token_accuracy": 0.7503726394497789, |
| "eval_num_tokens": 2026498.0, |
| "eval_runtime": 25.4801, |
| "eval_samples_per_second": 20.094, |
| "eval_steps_per_second": 10.047, |
| "step": 1000 |
| }, |
| { |
| "entropy": 1.9358359456062317, |
| "epoch": 2.731459532678632, |
| "grad_norm": 5.799363136291504, |
| "learning_rate": 9.020917811403104e-06, |
| "loss": 0.7895758152008057, |
| "mean_token_accuracy": 0.8083623085170984, |
| "num_tokens": 2045459.0, |
| "step": 1010 |
| }, |
| { |
| "entropy": 1.9362391978502274, |
| "epoch": 2.758550626481544, |
| "grad_norm": 4.842373371124268, |
| "learning_rate": 8.846801546080062e-06, |
| "loss": 0.836983585357666, |
| "mean_token_accuracy": 0.8126423776149749, |
| "num_tokens": 2066518.0, |
| "step": 1020 |
| }, |
| { |
| "entropy": 1.91400406062603, |
| "epoch": 2.7856417202844566, |
| "grad_norm": 7.656473636627197, |
| "learning_rate": 8.673038909714792e-06, |
| "loss": 0.7076562881469727, |
| "mean_token_accuracy": 0.8479138355702162, |
| "num_tokens": 2086183.0, |
| "step": 1030 |
| }, |
| { |
| "entropy": 1.8541285216808319, |
| "epoch": 2.8127328140873686, |
| "grad_norm": 9.124068260192871, |
| "learning_rate": 8.499683186717964e-06, |
| "loss": 0.9584396362304688, |
| "mean_token_accuracy": 0.7849915137514472, |
| "num_tokens": 2107721.0, |
| "step": 1040 |
| }, |
| { |
| "entropy": 1.798606288433075, |
| "epoch": 2.839823907890281, |
| "grad_norm": 11.562939643859863, |
| "learning_rate": 8.326787536720037e-06, |
| "loss": 0.9054560661315918, |
| "mean_token_accuracy": 0.7929567001760006, |
| "num_tokens": 2130459.0, |
| "step": 1050 |
| }, |
| { |
| "entropy": 1.9321310743689537, |
| "epoch": 2.8669150016931932, |
| "grad_norm": 12.520683288574219, |
| "learning_rate": 8.154404978269808e-06, |
| "loss": 0.9787432670593261, |
| "mean_token_accuracy": 0.7586101952940225, |
| "num_tokens": 2150465.0, |
| "step": 1060 |
| }, |
| { |
| "entropy": 1.8978336572647094, |
| "epoch": 2.8940060954961058, |
| "grad_norm": 6.707414627075195, |
| "learning_rate": 7.982588372576274e-06, |
| "loss": 0.8695780754089355, |
| "mean_token_accuracy": 0.8042063070461154, |
| "num_tokens": 2171901.0, |
| "step": 1070 |
| }, |
| { |
| "entropy": 1.911924734711647, |
| "epoch": 2.921097189299018, |
| "grad_norm": 10.575072288513184, |
| "learning_rate": 7.811390407298697e-06, |
| "loss": 1.1379961013793944, |
| "mean_token_accuracy": 0.7801183510571719, |
| "num_tokens": 2192878.0, |
| "step": 1080 |
| }, |
| { |
| "entropy": 1.9139499440789223, |
| "epoch": 2.9481882831019304, |
| "grad_norm": 9.588115692138672, |
| "learning_rate": 7.640863580389903e-06, |
| "loss": 0.6681862354278565, |
| "mean_token_accuracy": 0.8321586221456527, |
| "num_tokens": 2212872.0, |
| "step": 1090 |
| }, |
| { |
| "entropy": 1.917321428656578, |
| "epoch": 2.9752793769048425, |
| "grad_norm": 8.42562484741211, |
| "learning_rate": 7.47106018399776e-06, |
| "loss": 1.0313225746154786, |
| "mean_token_accuracy": 0.7394453713670373, |
| "num_tokens": 2233656.0, |
| "step": 1100 |
| }, |
| { |
| "epoch": 2.9752793769048425, |
| "eval_entropy": 1.8692557462491095, |
| "eval_loss": 1.1177661418914795, |
| "eval_mean_token_accuracy": 0.7488149127457291, |
| "eval_num_tokens": 2233656.0, |
| "eval_runtime": 25.4569, |
| "eval_samples_per_second": 20.112, |
| "eval_steps_per_second": 10.056, |
| "step": 1100 |
| }, |
| { |
| "entropy": 1.9485157532234714, |
| "epoch": 3.0, |
| "grad_norm": 20.587278366088867, |
| "learning_rate": 7.3020322884297565e-06, |
| "loss": 0.7162185668945312, |
| "mean_token_accuracy": 0.8130205739034365, |
| "num_tokens": 2251425.0, |
| "step": 1110 |
| }, |
| { |
| "entropy": 1.8808414414525032, |
| "epoch": 3.027091093802912, |
| "grad_norm": 9.604867935180664, |
| "learning_rate": 7.1338317261856225e-06, |
| "loss": 0.6537890911102295, |
| "mean_token_accuracy": 0.8606590256094933, |
| "num_tokens": 2270749.0, |
| "step": 1120 |
| }, |
| { |
| "entropy": 1.7552776366472245, |
| "epoch": 3.0541821876058246, |
| "grad_norm": 20.14252281188965, |
| "learning_rate": 6.966510076062845e-06, |
| "loss": 0.6271217823028564, |
| "mean_token_accuracy": 0.8392207693308592, |
| "num_tokens": 2289814.0, |
| "step": 1130 |
| }, |
| { |
| "entropy": 1.7204705789685248, |
| "epoch": 3.0812732814087367, |
| "grad_norm": 11.636978149414062, |
| "learning_rate": 6.800118647340022e-06, |
| "loss": 0.6068019866943359, |
| "mean_token_accuracy": 0.8635438393801451, |
| "num_tokens": 2310249.0, |
| "step": 1140 |
| }, |
| { |
| "entropy": 1.7508003085851669, |
| "epoch": 3.108364375211649, |
| "grad_norm": 12.074075698852539, |
| "learning_rate": 6.634708464042827e-06, |
| "loss": 0.43173890113830565, |
| "mean_token_accuracy": 0.8833881586790084, |
| "num_tokens": 2329569.0, |
| "step": 1150 |
| }, |
| { |
| "entropy": 1.7503153815865518, |
| "epoch": 3.1354554690145613, |
| "grad_norm": 12.822123527526855, |
| "learning_rate": 6.470330249297472e-06, |
| "loss": 0.5606254577636719, |
| "mean_token_accuracy": 0.8595870833843946, |
| "num_tokens": 2350193.0, |
| "step": 1160 |
| }, |
| { |
| "entropy": 1.7921119689941407, |
| "epoch": 3.162546562817474, |
| "grad_norm": 15.029194831848145, |
| "learning_rate": 6.3070344097764265e-06, |
| "loss": 0.775139856338501, |
| "mean_token_accuracy": 0.8431978467851877, |
| "num_tokens": 2369022.0, |
| "step": 1170 |
| }, |
| { |
| "entropy": 1.7634284302592278, |
| "epoch": 3.189637656620386, |
| "grad_norm": 10.424407958984375, |
| "learning_rate": 6.144871020241197e-06, |
| "loss": 0.6365249156951904, |
| "mean_token_accuracy": 0.8474900238215923, |
| "num_tokens": 2388527.0, |
| "step": 1180 |
| }, |
| { |
| "entropy": 1.7646004989743234, |
| "epoch": 3.2167287504232984, |
| "grad_norm": 8.618444442749023, |
| "learning_rate": 5.98388980818684e-06, |
| "loss": 0.8289416313171387, |
| "mean_token_accuracy": 0.8227736797183752, |
| "num_tokens": 2409237.0, |
| "step": 1190 |
| }, |
| { |
| "entropy": 1.737740470468998, |
| "epoch": 3.2438198442262105, |
| "grad_norm": 6.645010948181152, |
| "learning_rate": 5.8241401385930395e-06, |
| "loss": 0.39721031188964845, |
| "mean_token_accuracy": 0.8825947981327772, |
| "num_tokens": 2428916.0, |
| "step": 1200 |
| }, |
| { |
| "epoch": 3.2438198442262105, |
| "eval_entropy": 1.7245915541425347, |
| "eval_loss": 1.2651240825653076, |
| "eval_mean_token_accuracy": 0.7416992684593424, |
| "eval_num_tokens": 2428916.0, |
| "eval_runtime": 25.525, |
| "eval_samples_per_second": 20.059, |
| "eval_steps_per_second": 10.029, |
| "step": 1200 |
| }, |
| { |
| "entropy": 1.7320396453142166, |
| "epoch": 3.270910938029123, |
| "grad_norm": 9.127442359924316, |
| "learning_rate": 5.665670998786274e-06, |
| "loss": 0.5321141719818115, |
| "mean_token_accuracy": 0.8595851019024849, |
| "num_tokens": 2448191.0, |
| "step": 1210 |
| }, |
| { |
| "entropy": 1.7059780359268188, |
| "epoch": 3.298002031832035, |
| "grad_norm": 4.298802375793457, |
| "learning_rate": 5.508530983417855e-06, |
| "loss": 0.7181911468505859, |
| "mean_token_accuracy": 0.8558499291539192, |
| "num_tokens": 2470698.0, |
| "step": 1220 |
| }, |
| { |
| "entropy": 1.7368447184562683, |
| "epoch": 3.3250931256349476, |
| "grad_norm": 8.62894344329834, |
| "learning_rate": 5.352768279562315e-06, |
| "loss": 0.5110462188720704, |
| "mean_token_accuracy": 0.8599414020776749, |
| "num_tokens": 2490082.0, |
| "step": 1230 |
| }, |
| { |
| "entropy": 1.7677990421652794, |
| "epoch": 3.3521842194378597, |
| "grad_norm": 10.864715576171875, |
| "learning_rate": 5.198430651940846e-06, |
| "loss": 0.8702507972717285, |
| "mean_token_accuracy": 0.8074817329645156, |
| "num_tokens": 2510650.0, |
| "step": 1240 |
| }, |
| { |
| "entropy": 1.7419419810175896, |
| "epoch": 3.3792753132407722, |
| "grad_norm": 17.168798446655273, |
| "learning_rate": 5.0455654282742e-06, |
| "loss": 0.6099933624267578, |
| "mean_token_accuracy": 0.8637857645750046, |
| "num_tokens": 2530169.0, |
| "step": 1250 |
| }, |
| { |
| "entropy": 1.718581235408783, |
| "epoch": 3.4063664070436843, |
| "grad_norm": 9.985060691833496, |
| "learning_rate": 4.894219484769622e-06, |
| "loss": 0.38284480571746826, |
| "mean_token_accuracy": 0.894951194524765, |
| "num_tokens": 2549783.0, |
| "step": 1260 |
| }, |
| { |
| "entropy": 1.7342086032032966, |
| "epoch": 3.433457500846597, |
| "grad_norm": 7.34379768371582, |
| "learning_rate": 4.744439231746251e-06, |
| "loss": 0.5211464881896972, |
| "mean_token_accuracy": 0.8848932076245546, |
| "num_tokens": 2569240.0, |
| "step": 1270 |
| }, |
| { |
| "entropy": 1.7445754528045654, |
| "epoch": 3.460548594649509, |
| "grad_norm": 12.614356994628906, |
| "learning_rate": 4.596270599403338e-06, |
| "loss": 0.7350710391998291, |
| "mean_token_accuracy": 0.8188348516821862, |
| "num_tokens": 2589517.0, |
| "step": 1280 |
| }, |
| { |
| "entropy": 1.6525848761200905, |
| "epoch": 3.4876396884524214, |
| "grad_norm": 10.50085163116455, |
| "learning_rate": 4.449759023735749e-06, |
| "loss": 0.36133971214294436, |
| "mean_token_accuracy": 0.9022616557776928, |
| "num_tokens": 2610330.0, |
| "step": 1290 |
| }, |
| { |
| "entropy": 1.695315808057785, |
| "epoch": 3.5147307822553335, |
| "grad_norm": 7.3657026290893555, |
| "learning_rate": 4.30494943260098e-06, |
| "loss": 0.8237950325012207, |
| "mean_token_accuracy": 0.8609236305579543, |
| "num_tokens": 2632100.0, |
| "step": 1300 |
| }, |
| { |
| "epoch": 3.5147307822553335, |
| "eval_entropy": 1.7062555481679738, |
| "eval_loss": 1.265852451324463, |
| "eval_mean_token_accuracy": 0.7387107044924051, |
| "eval_num_tokens": 2632100.0, |
| "eval_runtime": 25.4553, |
| "eval_samples_per_second": 20.114, |
| "eval_steps_per_second": 10.057, |
| "step": 1300 |
| }, |
| { |
| "entropy": 1.7215346291661262, |
| "epoch": 3.541821876058246, |
| "grad_norm": 9.210244178771973, |
| "learning_rate": 4.161886231942017e-06, |
| "loss": 0.6864943027496337, |
| "mean_token_accuracy": 0.8675076253712177, |
| "num_tokens": 2652389.0, |
| "step": 1310 |
| }, |
| { |
| "entropy": 1.7494278416037559, |
| "epoch": 3.568912969861158, |
| "grad_norm": 12.154091835021973, |
| "learning_rate": 4.020613292170211e-06, |
| "loss": 0.672312593460083, |
| "mean_token_accuracy": 0.8510223753750324, |
| "num_tokens": 2672161.0, |
| "step": 1320 |
| }, |
| { |
| "entropy": 1.7348069936037063, |
| "epoch": 3.59600406366407, |
| "grad_norm": 10.369400978088379, |
| "learning_rate": 3.8811739347124e-06, |
| "loss": 0.4440812110900879, |
| "mean_token_accuracy": 0.8849797431379557, |
| "num_tokens": 2693004.0, |
| "step": 1330 |
| }, |
| { |
| "entropy": 1.721322275698185, |
| "epoch": 3.6230951574669827, |
| "grad_norm": 4.361479759216309, |
| "learning_rate": 3.7436109187263693e-06, |
| "loss": 0.52271409034729, |
| "mean_token_accuracy": 0.8836568117141723, |
| "num_tokens": 2712634.0, |
| "step": 1340 |
| }, |
| { |
| "entropy": 1.7014023318886757, |
| "epoch": 3.6501862512698953, |
| "grad_norm": 15.928149223327637, |
| "learning_rate": 3.6079664279887427e-06, |
| "loss": 0.7015122890472412, |
| "mean_token_accuracy": 0.8756893876940012, |
| "num_tokens": 2732861.0, |
| "step": 1350 |
| }, |
| { |
| "entropy": 1.7075004890561103, |
| "epoch": 3.6772773450728073, |
| "grad_norm": 14.396157264709473, |
| "learning_rate": 3.4742820579592673e-06, |
| "loss": 0.5177778720855712, |
| "mean_token_accuracy": 0.8753840968012809, |
| "num_tokens": 2752885.0, |
| "step": 1360 |
| }, |
| { |
| "entropy": 1.6834155157208444, |
| "epoch": 3.7043684388757194, |
| "grad_norm": 8.488550186157227, |
| "learning_rate": 3.342598803025595e-06, |
| "loss": 0.619942045211792, |
| "mean_token_accuracy": 0.8745716098695994, |
| "num_tokens": 2773651.0, |
| "step": 1370 |
| }, |
| { |
| "entropy": 1.7230923131108284, |
| "epoch": 3.731459532678632, |
| "grad_norm": 13.944185256958008, |
| "learning_rate": 3.212957043932283e-06, |
| "loss": 0.6659866809844971, |
| "mean_token_accuracy": 0.8391641702502965, |
| "num_tokens": 2794243.0, |
| "step": 1380 |
| }, |
| { |
| "entropy": 1.661913888156414, |
| "epoch": 3.758550626481544, |
| "grad_norm": 5.620434761047363, |
| "learning_rate": 3.085396535398044e-06, |
| "loss": 0.40467143058776855, |
| "mean_token_accuracy": 0.8812628626823426, |
| "num_tokens": 2815288.0, |
| "step": 1390 |
| }, |
| { |
| "entropy": 1.6734888523817062, |
| "epoch": 3.7856417202844566, |
| "grad_norm": 12.459346771240234, |
| "learning_rate": 2.959956393924922e-06, |
| "loss": 0.4842523574829102, |
| "mean_token_accuracy": 0.8642897168174386, |
| "num_tokens": 2835949.0, |
| "step": 1400 |
| }, |
| { |
| "epoch": 3.7856417202844566, |
| "eval_entropy": 1.6953813284635544, |
| "eval_loss": 1.3081457614898682, |
| "eval_mean_token_accuracy": 0.7404504243750125, |
| "eval_num_tokens": 2835949.0, |
| "eval_runtime": 25.4537, |
| "eval_samples_per_second": 20.115, |
| "eval_steps_per_second": 10.057, |
| "step": 1400 |
| }, |
| { |
| "entropy": 1.7142169624567032, |
| "epoch": 3.8127328140873686, |
| "grad_norm": 19.3244571685791, |
| "learning_rate": 2.8366750858032177e-06, |
| "loss": 0.7156589031219482, |
| "mean_token_accuracy": 0.8869716469198465, |
| "num_tokens": 2856246.0, |
| "step": 1410 |
| }, |
| { |
| "entropy": 1.6850622862577438, |
| "epoch": 3.839823907890281, |
| "grad_norm": 15.101435661315918, |
| "learning_rate": 2.7155904153157497e-06, |
| "loss": 0.45338053703308107, |
| "mean_token_accuracy": 0.8818465910851956, |
| "num_tokens": 2877330.0, |
| "step": 1420 |
| }, |
| { |
| "entropy": 1.7075368210673332, |
| "epoch": 3.8669150016931932, |
| "grad_norm": 15.861879348754883, |
| "learning_rate": 2.5967395131451723e-06, |
| "loss": 0.5139744758605957, |
| "mean_token_accuracy": 0.8527137745171786, |
| "num_tokens": 2897344.0, |
| "step": 1430 |
| }, |
| { |
| "entropy": 1.6651010170578957, |
| "epoch": 3.8940060954961058, |
| "grad_norm": 15.302207946777344, |
| "learning_rate": 2.4801588249878163e-06, |
| "loss": 0.7968846797943115, |
| "mean_token_accuracy": 0.8468474011868239, |
| "num_tokens": 2919143.0, |
| "step": 1440 |
| }, |
| { |
| "entropy": 1.6404287233948707, |
| "epoch": 3.921097189299018, |
| "grad_norm": 10.076866149902344, |
| "learning_rate": 2.365884100377611e-06, |
| "loss": 0.4457272529602051, |
| "mean_token_accuracy": 0.8820912927389145, |
| "num_tokens": 2941599.0, |
| "step": 1450 |
| }, |
| { |
| "entropy": 1.6921194523572922, |
| "epoch": 3.9481882831019304, |
| "grad_norm": 10.407143592834473, |
| "learning_rate": 2.2539503817234553e-06, |
| "loss": 0.6064411640167237, |
| "mean_token_accuracy": 0.8608082607388496, |
| "num_tokens": 2961983.0, |
| "step": 1460 |
| }, |
| { |
| "entropy": 1.712096980214119, |
| "epoch": 3.9752793769048425, |
| "grad_norm": 10.601531028747559, |
| "learning_rate": 2.144391993563446e-06, |
| "loss": 0.503159761428833, |
| "mean_token_accuracy": 0.8580688439309597, |
| "num_tokens": 2982164.0, |
| "step": 1470 |
| }, |
| { |
| "entropy": 1.6592551649433294, |
| "epoch": 4.0, |
| "grad_norm": 21.362045288085938, |
| "learning_rate": 2.0372425320392454e-06, |
| "loss": 0.5836192607879639, |
| "mean_token_accuracy": 0.8683291808383106, |
| "num_tokens": 3001900.0, |
| "step": 1480 |
| }, |
| { |
| "entropy": 1.6971124514937401, |
| "epoch": 4.0270910938029125, |
| "grad_norm": 8.777483940124512, |
| "learning_rate": 1.932534854593795e-06, |
| "loss": 0.6050891876220703, |
| "mean_token_accuracy": 0.8690009146928788, |
| "num_tokens": 3023056.0, |
| "step": 1490 |
| }, |
| { |
| "entropy": 1.63857099711895, |
| "epoch": 4.054182187605824, |
| "grad_norm": 8.595149040222168, |
| "learning_rate": 1.8303010698955803e-06, |
| "loss": 0.4310460090637207, |
| "mean_token_accuracy": 0.8924087427556515, |
| "num_tokens": 3043490.0, |
| "step": 1500 |
| }, |
| { |
| "epoch": 4.054182187605824, |
| "eval_entropy": 1.669337394181639, |
| "eval_loss": 1.3523277044296265, |
| "eval_mean_token_accuracy": 0.7376706907525659, |
| "eval_num_tokens": 3043490.0, |
| "eval_runtime": 25.4594, |
| "eval_samples_per_second": 20.11, |
| "eval_steps_per_second": 10.055, |
| "step": 1500 |
| }, |
| { |
| "entropy": 1.6625288650393486, |
| "epoch": 4.081273281408737, |
| "grad_norm": 12.073065757751465, |
| "learning_rate": 1.7305725279924634e-06, |
| "loss": 0.3827295541763306, |
| "mean_token_accuracy": 0.8844121858477593, |
| "num_tokens": 3064126.0, |
| "step": 1510 |
| }, |
| { |
| "entropy": 1.6116922572255135, |
| "epoch": 4.108364375211649, |
| "grad_norm": 5.359347820281982, |
| "learning_rate": 1.6333798106982024e-06, |
| "loss": 0.3580256700515747, |
| "mean_token_accuracy": 0.9377022050321102, |
| "num_tokens": 3083867.0, |
| "step": 1520 |
| }, |
| { |
| "entropy": 1.591383346915245, |
| "epoch": 4.135455469014562, |
| "grad_norm": 6.666597843170166, |
| "learning_rate": 1.5387527222144861e-06, |
| "loss": 0.43925886154174804, |
| "mean_token_accuracy": 0.918615211546421, |
| "num_tokens": 3103679.0, |
| "step": 1530 |
| }, |
| { |
| "entropy": 1.5715536609292031, |
| "epoch": 4.162546562817473, |
| "grad_norm": 9.182818412780762, |
| "learning_rate": 1.446720279991496e-06, |
| "loss": 0.30636699199676515, |
| "mean_token_accuracy": 0.9238971933722496, |
| "num_tokens": 3125843.0, |
| "step": 1540 |
| }, |
| { |
| "entropy": 1.6449363514781, |
| "epoch": 4.189637656620386, |
| "grad_norm": 7.822624206542969, |
| "learning_rate": 1.3573107058296619e-06, |
| "loss": 0.45112295150756837, |
| "mean_token_accuracy": 0.8905558969825507, |
| "num_tokens": 3144837.0, |
| "step": 1550 |
| }, |
| { |
| "entropy": 1.5784786254167558, |
| "epoch": 4.216728750423298, |
| "grad_norm": 11.84586238861084, |
| "learning_rate": 1.270551417225443e-06, |
| "loss": 0.4005408763885498, |
| "mean_token_accuracy": 0.9311332777142525, |
| "num_tokens": 3163882.0, |
| "step": 1560 |
| }, |
| { |
| "entropy": 1.5929084509611129, |
| "epoch": 4.243819844226211, |
| "grad_norm": 18.798683166503906, |
| "learning_rate": 1.1864690189637306e-06, |
| "loss": 0.5499368190765381, |
| "mean_token_accuracy": 0.8895844966173172, |
| "num_tokens": 3185313.0, |
| "step": 1570 |
| }, |
| { |
| "entropy": 1.6276163026690482, |
| "epoch": 4.270910938029123, |
| "grad_norm": 18.325300216674805, |
| "learning_rate": 1.105089294959487e-06, |
| "loss": 0.47011642456054686, |
| "mean_token_accuracy": 0.8977843724191189, |
| "num_tokens": 3206038.0, |
| "step": 1580 |
| }, |
| { |
| "entropy": 1.6390855327248572, |
| "epoch": 4.298002031832035, |
| "grad_norm": 16.15447235107422, |
| "learning_rate": 1.0264372003510747e-06, |
| "loss": 0.5724862575531006, |
| "mean_token_accuracy": 0.8626310247927904, |
| "num_tokens": 3225061.0, |
| "step": 1590 |
| }, |
| { |
| "entropy": 1.6146731659770013, |
| "epoch": 4.325093125634948, |
| "grad_norm": 5.743528366088867, |
| "learning_rate": 9.5053685384777e-07, |
| "loss": 0.6115519046783447, |
| "mean_token_accuracy": 0.8991932325065136, |
| "num_tokens": 3246376.0, |
| "step": 1600 |
| }, |
| { |
| "epoch": 4.325093125634948, |
| "eval_entropy": 1.6513289087451994, |
| "eval_loss": 1.399103045463562, |
| "eval_mean_token_accuracy": 0.7364511488704011, |
| "eval_num_tokens": 3246376.0, |
| "eval_runtime": 25.4867, |
| "eval_samples_per_second": 20.089, |
| "eval_steps_per_second": 10.044, |
| "step": 1600 |
| }, |
| { |
| "entropy": 1.6438001811504364, |
| "epoch": 4.35218421943786, |
| "grad_norm": 10.316023826599121, |
| "learning_rate": 8.774115303337305e-07, |
| "loss": 0.3678164005279541, |
| "mean_token_accuracy": 0.9086852367967367, |
| "num_tokens": 3266626.0, |
| "step": 1610 |
| }, |
| { |
| "entropy": 1.6436322793364524, |
| "epoch": 4.379275313240772, |
| "grad_norm": 7.732473373413086, |
| "learning_rate": 8.070836537307536e-07, |
| "loss": 0.38886566162109376, |
| "mean_token_accuracy": 0.9289665892720222, |
| "num_tokens": 3285791.0, |
| "step": 1620 |
| }, |
| { |
| "entropy": 1.6396998748183251, |
| "epoch": 4.406366407043684, |
| "grad_norm": 15.49699592590332, |
| "learning_rate": 7.395747901219474e-07, |
| "loss": 0.3827892541885376, |
| "mean_token_accuracy": 0.9125935144722461, |
| "num_tokens": 3304914.0, |
| "step": 1630 |
| }, |
| { |
| "entropy": 1.6532684773206712, |
| "epoch": 4.433457500846597, |
| "grad_norm": 14.576912879943848, |
| "learning_rate": 6.749056411385046e-07, |
| "loss": 0.6813682079315185, |
| "mean_token_accuracy": 0.8684317111968994, |
| "num_tokens": 3324646.0, |
| "step": 1640 |
| }, |
| { |
| "entropy": 1.6200390428304672, |
| "epoch": 4.460548594649509, |
| "grad_norm": 18.16828727722168, |
| "learning_rate": 6.130960376115147e-07, |
| "loss": 0.38154261112213134, |
| "mean_token_accuracy": 0.8886280782520771, |
| "num_tokens": 3345568.0, |
| "step": 1650 |
| }, |
| { |
| "entropy": 1.6201925188302995, |
| "epoch": 4.487639688452421, |
| "grad_norm": 4.342578887939453, |
| "learning_rate": 5.541649334908405e-07, |
| "loss": 0.3788790464401245, |
| "mean_token_accuracy": 0.891268914937973, |
| "num_tokens": 3366108.0, |
| "step": 1660 |
| }, |
| { |
| "entropy": 1.6324064388871193, |
| "epoch": 4.5147307822553335, |
| "grad_norm": 10.33731460571289, |
| "learning_rate": 4.981304000328702e-07, |
| "loss": 0.491044807434082, |
| "mean_token_accuracy": 0.9079089991748333, |
| "num_tokens": 3386012.0, |
| "step": 1670 |
| }, |
| { |
| "entropy": 1.613644614815712, |
| "epoch": 4.541821876058246, |
| "grad_norm": 8.337615966796875, |
| "learning_rate": 4.450096202589671e-07, |
| "loss": 0.28208341598510744, |
| "mean_token_accuracy": 0.9069402247667313, |
| "num_tokens": 3406651.0, |
| "step": 1680 |
| }, |
| { |
| "entropy": 1.585978364944458, |
| "epoch": 4.568912969861158, |
| "grad_norm": 8.048723220825195, |
| "learning_rate": 3.9481888368627764e-07, |
| "loss": 0.27262258529663086, |
| "mean_token_accuracy": 0.9263891167938709, |
| "num_tokens": 3426691.0, |
| "step": 1690 |
| }, |
| { |
| "entropy": 1.6033645749092102, |
| "epoch": 4.59600406366407, |
| "grad_norm": 15.038307189941406, |
| "learning_rate": 3.4757358133254207e-07, |
| "loss": 0.6044834136962891, |
| "mean_token_accuracy": 0.8724760405719281, |
| "num_tokens": 3447608.0, |
| "step": 1700 |
| }, |
| { |
| "epoch": 4.59600406366407, |
| "eval_entropy": 1.645897411275655, |
| "eval_loss": 1.4024633169174194, |
| "eval_mean_token_accuracy": 0.7368950636591762, |
| "eval_num_tokens": 3447608.0, |
| "eval_runtime": 25.5248, |
| "eval_samples_per_second": 20.059, |
| "eval_steps_per_second": 10.029, |
| "step": 1700 |
| }, |
| { |
| "entropy": 1.6228346675634384, |
| "epoch": 4.623095157466983, |
| "grad_norm": 13.567054748535156, |
| "learning_rate": 3.032882009964322e-07, |
| "loss": 0.4227153301239014, |
| "mean_token_accuracy": 0.9056596923619509, |
| "num_tokens": 3467735.0, |
| "step": 1710 |
| }, |
| { |
| "entropy": 1.617300546169281, |
| "epoch": 4.650186251269895, |
| "grad_norm": 14.03318977355957, |
| "learning_rate": 2.619763228148664e-07, |
| "loss": 0.35500648021698, |
| "mean_token_accuracy": 0.9040526457130909, |
| "num_tokens": 3487598.0, |
| "step": 1720 |
| }, |
| { |
| "entropy": 1.5993121460080146, |
| "epoch": 4.677277345072808, |
| "grad_norm": 9.692215919494629, |
| "learning_rate": 2.236506150986395e-07, |
| "loss": 0.44063520431518555, |
| "mean_token_accuracy": 0.9209048740565777, |
| "num_tokens": 3508359.0, |
| "step": 1730 |
| }, |
| { |
| "entropy": 1.6759715333580971, |
| "epoch": 4.704368438875719, |
| "grad_norm": 19.903579711914062, |
| "learning_rate": 1.8832283044768585e-07, |
| "loss": 0.4889723300933838, |
| "mean_token_accuracy": 0.8778417184948921, |
| "num_tokens": 3527446.0, |
| "step": 1740 |
| }, |
| { |
| "entropy": 1.6152383774518966, |
| "epoch": 4.731459532678632, |
| "grad_norm": 3.170473337173462, |
| "learning_rate": 1.5600380214714216e-07, |
| "loss": 0.4727470397949219, |
| "mean_token_accuracy": 0.9017321106046439, |
| "num_tokens": 3547726.0, |
| "step": 1750 |
| }, |
| { |
| "entropy": 1.6021659523248672, |
| "epoch": 4.7585506264815445, |
| "grad_norm": 20.662057876586914, |
| "learning_rate": 1.2670344084530384e-07, |
| "loss": 0.46443953514099123, |
| "mean_token_accuracy": 0.8934749307110905, |
| "num_tokens": 3567854.0, |
| "step": 1760 |
| }, |
| { |
| "entropy": 1.5918526247143745, |
| "epoch": 4.785641720284456, |
| "grad_norm": 10.510283470153809, |
| "learning_rate": 1.0043073151452808e-07, |
| "loss": 0.4407022476196289, |
| "mean_token_accuracy": 0.9137253064662219, |
| "num_tokens": 3589923.0, |
| "step": 1770 |
| }, |
| { |
| "entropy": 1.6411705940961838, |
| "epoch": 4.812732814087369, |
| "grad_norm": 15.203192710876465, |
| "learning_rate": 7.719373069598246e-08, |
| "loss": 0.49732284545898436, |
| "mean_token_accuracy": 0.8905138060450554, |
| "num_tokens": 3610920.0, |
| "step": 1780 |
| }, |
| { |
| "entropy": 1.623388308286667, |
| "epoch": 4.839823907890281, |
| "grad_norm": 12.377976417541504, |
| "learning_rate": 5.6999564029103226e-08, |
| "loss": 0.4741045475006104, |
| "mean_token_accuracy": 0.9054140999913216, |
| "num_tokens": 3630913.0, |
| "step": 1790 |
| }, |
| { |
| "entropy": 1.5931948989629745, |
| "epoch": 4.866915001693194, |
| "grad_norm": 7.101802825927734, |
| "learning_rate": 3.98544240665133e-08, |
| "loss": 0.4587404251098633, |
| "mean_token_accuracy": 0.9269338101148605, |
| "num_tokens": 3651374.0, |
| "step": 1800 |
| }, |
| { |
| "epoch": 4.866915001693194, |
| "eval_entropy": 1.6439465275034308, |
| "eval_loss": 1.4104795455932617, |
| "eval_mean_token_accuracy": 0.736133792786859, |
| "eval_num_tokens": 3651374.0, |
| "eval_runtime": 25.4587, |
| "eval_samples_per_second": 20.111, |
| "eval_steps_per_second": 10.055, |
| "step": 1800 |
| }, |
| { |
| "entropy": 1.601736557483673, |
| "epoch": 4.894006095496105, |
| "grad_norm": 11.146471977233887, |
| "learning_rate": 2.5763568375075655e-08, |
| "loss": 0.5720036029815674, |
| "mean_token_accuracy": 0.8911033194512129, |
| "num_tokens": 3671445.0, |
| "step": 1810 |
| }, |
| { |
| "entropy": 1.60008362531662, |
| "epoch": 4.921097189299018, |
| "grad_norm": 21.521833419799805, |
| "learning_rate": 1.473131792365301e-08, |
| "loss": 0.37158544063568116, |
| "mean_token_accuracy": 0.9116154242306947, |
| "num_tokens": 3691881.0, |
| "step": 1820 |
| }, |
| { |
| "entropy": 1.6304287910461426, |
| "epoch": 4.94818828310193, |
| "grad_norm": 15.239692687988281, |
| "learning_rate": 6.761055758083279e-09, |
| "loss": 0.42179555892944337, |
| "mean_token_accuracy": 0.8883754149079323, |
| "num_tokens": 3711706.0, |
| "step": 1830 |
| }, |
| { |
| "entropy": 1.6515601187944413, |
| "epoch": 4.975279376904843, |
| "grad_norm": 11.510629653930664, |
| "learning_rate": 1.8552259637627524e-09, |
| "loss": 0.3710385799407959, |
| "mean_token_accuracy": 0.9058502331376076, |
| "num_tokens": 3733729.0, |
| "step": 1840 |
| }, |
| { |
| "entropy": 1.6298474122400153, |
| "epoch": 5.0, |
| "grad_norm": 8.197460174560547, |
| "learning_rate": 1.5332916172283718e-11, |
| "loss": 0.3809305429458618, |
| "mean_token_accuracy": 0.9352603583303216, |
| "num_tokens": 3752375.0, |
| "step": 1850 |
| }, |
| { |
| "epoch": 5.0, |
| "eval_entropy": 1.6423234520480037, |
| "eval_loss": 1.4100017547607422, |
| "eval_mean_token_accuracy": 0.7342472999589518, |
| "eval_num_tokens": 3752375.0, |
| "eval_runtime": 25.4872, |
| "eval_samples_per_second": 20.089, |
| "eval_steps_per_second": 10.044, |
| "step": 1850 |
| }, |
| { |
| "epoch": 5.0, |
| "step": 1850, |
| "total_flos": 1.69605163656192e+17, |
| "train_loss": 0.9120298720694877, |
| "train_runtime": 3361.88, |
| "train_samples_per_second": 4.392, |
| "train_steps_per_second": 0.55 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 1850, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 5, |
| "save_steps": 100, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.69605163656192e+17, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
