trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 8.012820512820513,
  "eval_steps": 500,
  "global_step": 15000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2670940170940171,
      "grad_norm": 0.11774194240570068,
      "learning_rate": 4.9809218559218565e-05,
      "loss": 0.246,
      "step": 500
    },
    {
      "epoch": 0.2670940170940171,
      "eval_f1": 0.5214187643020595,
      "eval_loss": 0.03276558592915535,
      "eval_precision": 0.5067158868528732,
      "eval_recall": 0.5370003770739065,
      "eval_runtime": 119.6438,
      "eval_samples_per_second": 123.349,
      "eval_steps_per_second": 3.861,
      "step": 500
    },
    {
      "epoch": 0.5341880341880342,
      "grad_norm": 0.042453471571207047,
      "learning_rate": 4.961843711843712e-05,
      "loss": 0.0238,
      "step": 1000
    },
    {
      "epoch": 0.5341880341880342,
      "eval_f1": 0.6635963067112459,
      "eval_loss": 0.01632688380777836,
      "eval_precision": 0.6257568797761723,
      "eval_recall": 0.7063065610859729,
      "eval_runtime": 125.2059,
      "eval_samples_per_second": 117.87,
      "eval_steps_per_second": 3.69,
      "step": 1000
    },
    {
      "epoch": 0.8012820512820513,
      "grad_norm": 0.024465590715408325,
      "learning_rate": 4.9427655677655683e-05,
      "loss": 0.0127,
      "step": 1500
    },
    {
      "epoch": 0.8012820512820513,
      "eval_f1": 0.9185966204705558,
      "eval_loss": 0.007961403578519821,
      "eval_precision": 0.8978039780397804,
      "eval_recall": 0.9403751885369532,
      "eval_runtime": 214.5953,
      "eval_samples_per_second": 68.771,
      "eval_steps_per_second": 2.153,
      "step": 1500
    },
    {
      "epoch": 1.0683760683760684,
      "grad_norm": 0.06622592359781265,
      "learning_rate": 4.923687423687424e-05,
      "loss": 0.0063,
      "step": 2000
    },
    {
      "epoch": 1.0683760683760684,
      "eval_f1": 0.9344357210681118,
      "eval_loss": 0.005176798906177282,
      "eval_precision": 0.9185448253881529,
      "eval_recall": 0.9508861236802413,
      "eval_runtime": 215.2423,
      "eval_samples_per_second": 68.565,
      "eval_steps_per_second": 2.146,
      "step": 2000
    },
    {
      "epoch": 1.3354700854700854,
      "grad_norm": 0.048759885132312775,
      "learning_rate": 4.9046092796092795e-05,
      "loss": 0.0041,
      "step": 2500
    },
    {
      "epoch": 1.3354700854700854,
      "eval_f1": 0.9516508609641868,
      "eval_loss": 0.0035495704505592585,
      "eval_precision": 0.9410165430164509,
      "eval_recall": 0.9625282805429864,
      "eval_runtime": 215.3598,
      "eval_samples_per_second": 68.527,
      "eval_steps_per_second": 2.145,
      "step": 2500
    },
    {
      "epoch": 1.6025641025641026,
      "grad_norm": 0.04741215705871582,
      "learning_rate": 4.885531135531136e-05,
      "loss": 0.003,
      "step": 3000
    },
    {
      "epoch": 1.6025641025641026,
      "eval_f1": 0.9602090789225843,
      "eval_loss": 0.002707522828131914,
      "eval_precision": 0.9593056078283778,
      "eval_recall": 0.9611142533936652,
      "eval_runtime": 215.7177,
      "eval_samples_per_second": 68.413,
      "eval_steps_per_second": 2.142,
      "step": 3000
    },
    {
      "epoch": 1.8696581196581197,
      "grad_norm": 0.022911719977855682,
      "learning_rate": 4.8664529914529914e-05,
      "loss": 0.0024,
      "step": 3500
    },
    {
      "epoch": 1.8696581196581197,
      "eval_f1": 0.9571561338289963,
      "eval_loss": 0.0023886554408818483,
      "eval_precision": 0.9438233137829912,
      "eval_recall": 0.9708710407239819,
      "eval_runtime": 205.5029,
      "eval_samples_per_second": 71.814,
      "eval_steps_per_second": 2.248,
      "step": 3500
    },
    {
      "epoch": 2.1367521367521367,
      "grad_norm": 0.05229872837662697,
      "learning_rate": 4.8473748473748476e-05,
      "loss": 0.0018,
      "step": 4000
    },
    {
      "epoch": 2.1367521367521367,
      "eval_f1": 0.9639815719670529,
      "eval_loss": 0.0022004263009876013,
      "eval_precision": 0.9518886131789358,
      "eval_recall": 0.9763857466063348,
      "eval_runtime": 217.5053,
      "eval_samples_per_second": 67.851,
      "eval_steps_per_second": 2.124,
      "step": 4000
    },
    {
      "epoch": 2.4038461538461537,
      "grad_norm": 0.04928997904062271,
      "learning_rate": 4.828296703296704e-05,
      "loss": 0.0013,
      "step": 4500
    },
    {
      "epoch": 2.4038461538461537,
      "eval_f1": 0.9693363629974934,
      "eval_loss": 0.0017703943885862827,
      "eval_precision": 0.9635356028500908,
      "eval_recall": 0.9752073906485671,
      "eval_runtime": 205.3479,
      "eval_samples_per_second": 71.868,
      "eval_steps_per_second": 2.25,
      "step": 4500
    },
    {
      "epoch": 2.6709401709401708,
      "grad_norm": 0.03343936800956726,
      "learning_rate": 4.8092185592185595e-05,
      "loss": 0.0014,
      "step": 5000
    },
    {
      "epoch": 2.6709401709401708,
      "eval_f1": 0.9706270005648653,
      "eval_loss": 0.001739403698593378,
      "eval_precision": 0.9693493794659647,
      "eval_recall": 0.9719079939668175,
      "eval_runtime": 204.7418,
      "eval_samples_per_second": 72.081,
      "eval_steps_per_second": 2.257,
      "step": 5000
    },
    {
      "epoch": 2.9380341880341883,
      "grad_norm": 0.07879549264907837,
      "learning_rate": 4.790140415140416e-05,
      "loss": 0.0013,
      "step": 5500
    },
    {
      "epoch": 2.9380341880341883,
      "eval_f1": 0.9702524596250233,
      "eval_loss": 0.0016117716440930963,
      "eval_precision": 0.9555301645338209,
      "eval_recall": 0.985435520361991,
      "eval_runtime": 216.8331,
      "eval_samples_per_second": 68.062,
      "eval_steps_per_second": 2.131,
      "step": 5500
    },
    {
      "epoch": 3.2051282051282053,
      "grad_norm": 0.14366598427295685,
      "learning_rate": 4.771062271062271e-05,
      "loss": 0.0008,
      "step": 6000
    },
    {
      "epoch": 3.2051282051282053,
      "eval_f1": 0.977527028941032,
      "eval_loss": 0.001418485539034009,
      "eval_precision": 0.9791449919606545,
      "eval_recall": 0.9759144042232277,
      "eval_runtime": 217.6653,
      "eval_samples_per_second": 67.801,
      "eval_steps_per_second": 2.123,
      "step": 6000
    },
    {
      "epoch": 3.4722222222222223,
      "grad_norm": 0.01724930852651596,
      "learning_rate": 4.751984126984127e-05,
      "loss": 0.0009,
      "step": 6500
    },
    {
      "epoch": 3.4722222222222223,
      "eval_f1": 0.9748389220712035,
      "eval_loss": 0.0014127923641353846,
      "eval_precision": 0.9726888784608165,
      "eval_recall": 0.976998491704374,
      "eval_runtime": 120.5351,
      "eval_samples_per_second": 122.437,
      "eval_steps_per_second": 3.833,
      "step": 6500
    },
    {
      "epoch": 3.7393162393162394,
      "grad_norm": 0.0373033843934536,
      "learning_rate": 4.732905982905983e-05,
      "loss": 0.0008,
      "step": 7000
    },
    {
      "epoch": 3.7393162393162394,
      "eval_f1": 0.9767944381811348,
      "eval_loss": 0.001502642291598022,
      "eval_precision": 0.973501872659176,
      "eval_recall": 0.9801093514328808,
      "eval_runtime": 119.8163,
      "eval_samples_per_second": 123.172,
      "eval_steps_per_second": 3.856,
      "step": 7000
    },
    {
      "epoch": 4.006410256410256,
      "grad_norm": 0.004327626433223486,
      "learning_rate": 4.713827838827839e-05,
      "loss": 0.0007,
      "step": 7500
    },
    {
      "epoch": 4.006410256410256,
      "eval_f1": 0.9779735682819383,
      "eval_loss": 0.001390722580254078,
      "eval_precision": 0.9724137931034482,
      "eval_recall": 0.9835972850678733,
      "eval_runtime": 119.5661,
      "eval_samples_per_second": 123.43,
      "eval_steps_per_second": 3.864,
      "step": 7500
    },
    {
      "epoch": 4.273504273504273,
      "grad_norm": 0.0008369502611458302,
      "learning_rate": 4.694749694749695e-05,
      "loss": 0.0006,
      "step": 8000
    },
    {
      "epoch": 4.273504273504273,
      "eval_f1": 0.9788639365918097,
      "eval_loss": 0.0014948239549994469,
      "eval_precision": 0.9797884397431054,
      "eval_recall": 0.9779411764705882,
      "eval_runtime": 121.1034,
      "eval_samples_per_second": 121.863,
      "eval_steps_per_second": 3.815,
      "step": 8000
    },
    {
      "epoch": 4.5405982905982905,
      "grad_norm": 0.034645263105630875,
      "learning_rate": 4.675671550671551e-05,
      "loss": 0.0005,
      "step": 8500
    },
    {
      "epoch": 4.5405982905982905,
      "eval_f1": 0.9742678851052889,
      "eval_loss": 0.0017295886063948274,
      "eval_precision": 0.9662509851189096,
      "eval_recall": 0.9824189291101055,
      "eval_runtime": 214.7422,
      "eval_samples_per_second": 68.724,
      "eval_steps_per_second": 2.151,
      "step": 8500
    },
    {
      "epoch": 4.8076923076923075,
      "grad_norm": 0.010439756326377392,
      "learning_rate": 4.656593406593407e-05,
      "loss": 0.0006,
      "step": 9000
    },
    {
      "epoch": 4.8076923076923075,
      "eval_f1": 0.9778079816277271,
      "eval_loss": 0.0013649952597916126,
      "eval_precision": 0.9723167264762083,
      "eval_recall": 0.9833616138763197,
      "eval_runtime": 215.7325,
      "eval_samples_per_second": 68.409,
      "eval_steps_per_second": 2.142,
      "step": 9000
    },
    {
      "epoch": 5.0747863247863245,
      "grad_norm": 0.022446786984801292,
      "learning_rate": 4.637515262515263e-05,
      "loss": 0.0005,
      "step": 9500
    },
    {
      "epoch": 5.0747863247863245,
      "eval_f1": 0.9819268602626252,
      "eval_loss": 0.0013456182787194848,
      "eval_precision": 0.9804962872450418,
      "eval_recall": 0.9833616138763197,
      "eval_runtime": 145.9782,
      "eval_samples_per_second": 101.097,
      "eval_steps_per_second": 3.165,
      "step": 9500
    },
    {
      "epoch": 5.3418803418803416,
      "grad_norm": 0.0010687178000807762,
      "learning_rate": 4.618437118437119e-05,
      "loss": 0.0004,
      "step": 10000
    },
    {
      "epoch": 5.3418803418803416,
      "eval_f1": 0.9747765825886088,
      "eval_loss": 0.0019718140829354525,
      "eval_precision": 0.9652049350769373,
      "eval_recall": 0.9845399698340875,
      "eval_runtime": 218.776,
      "eval_samples_per_second": 67.457,
      "eval_steps_per_second": 2.112,
      "step": 10000
    },
    {
      "epoch": 5.608974358974359,
      "grad_norm": 0.010968877002596855,
      "learning_rate": 4.599358974358974e-05,
      "loss": 0.0005,
      "step": 10500
    },
    {
      "epoch": 5.608974358974359,
      "eval_f1": 0.9817333301906805,
      "eval_loss": 0.0013046471867710352,
      "eval_precision": 0.9818490405921456,
      "eval_recall": 0.9816176470588235,
      "eval_runtime": 205.7951,
      "eval_samples_per_second": 71.712,
      "eval_steps_per_second": 2.245,
      "step": 10500
    },
    {
      "epoch": 5.8760683760683765,
      "grad_norm": 0.0015924246981739998,
      "learning_rate": 4.5802808302808305e-05,
      "loss": 0.0004,
      "step": 11000
    },
    {
      "epoch": 5.8760683760683765,
      "eval_f1": 0.9823409981585532,
      "eval_loss": 0.0014677270082756877,
      "eval_precision": 0.9840601646012677,
      "eval_recall": 0.9806278280542986,
      "eval_runtime": 203.592,
      "eval_samples_per_second": 72.488,
      "eval_steps_per_second": 2.269,
      "step": 11000
    },
    {
      "epoch": 6.1431623931623935,
      "grad_norm": 0.004735188093036413,
      "learning_rate": 4.561202686202686e-05,
      "loss": 0.0004,
      "step": 11500
    },
    {
      "epoch": 6.1431623931623935,
      "eval_f1": 0.9827679432023886,
      "eval_loss": 0.0012729066656902432,
      "eval_precision": 0.9803480136954177,
      "eval_recall": 0.9851998491704375,
      "eval_runtime": 119.9744,
      "eval_samples_per_second": 123.01,
      "eval_steps_per_second": 3.851,
      "step": 11500
    },
    {
      "epoch": 6.410256410256411,
      "grad_norm": 0.008303467184305191,
      "learning_rate": 4.5421245421245424e-05,
      "loss": 0.0003,
      "step": 12000
    },
    {
      "epoch": 6.410256410256411,
      "eval_f1": 0.9750854909623841,
      "eval_loss": 0.0016973286401480436,
      "eval_precision": 0.9626567452115199,
      "eval_recall": 0.9878393665158371,
      "eval_runtime": 219.3026,
      "eval_samples_per_second": 67.295,
      "eval_steps_per_second": 2.107,
      "step": 12000
    },
    {
      "epoch": 6.677350427350428,
      "grad_norm": 0.007308937143534422,
      "learning_rate": 4.5230463980463986e-05,
      "loss": 0.0003,
      "step": 12500
    },
    {
      "epoch": 6.677350427350428,
      "eval_f1": 0.9813210991994756,
      "eval_loss": 0.0012571236584335566,
      "eval_precision": 0.9747047335627267,
      "eval_recall": 0.98802790346908,
      "eval_runtime": 219.2644,
      "eval_samples_per_second": 67.307,
      "eval_steps_per_second": 2.107,
      "step": 12500
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 0.05341866984963417,
      "learning_rate": 4.503968253968254e-05,
      "loss": 0.0004,
      "step": 13000
    },
    {
      "epoch": 6.944444444444445,
      "eval_f1": 0.9821235357764502,
      "eval_loss": 0.0013370579108595848,
      "eval_precision": 0.9802328857169688,
      "eval_recall": 0.9840214932126696,
      "eval_runtime": 219.0265,
      "eval_samples_per_second": 67.38,
      "eval_steps_per_second": 2.109,
      "step": 13000
    },
    {
      "epoch": 7.211538461538462,
      "grad_norm": 0.0043980698101222515,
      "learning_rate": 4.4848901098901105e-05,
      "loss": 0.0003,
      "step": 13500
    },
    {
      "epoch": 7.211538461538462,
      "eval_f1": 0.9812013269651554,
      "eval_loss": 0.00139136821962893,
      "eval_precision": 0.9795649927185607,
      "eval_recall": 0.9828431372549019,
      "eval_runtime": 205.0799,
      "eval_samples_per_second": 71.962,
      "eval_steps_per_second": 2.253,
      "step": 13500
    },
    {
      "epoch": 7.478632478632479,
      "grad_norm": 0.012946384958922863,
      "learning_rate": 4.465811965811966e-05,
      "loss": 0.0003,
      "step": 14000
    },
    {
      "epoch": 7.478632478632479,
      "eval_f1": 0.9811170275151885,
      "eval_loss": 0.0012917759595438838,
      "eval_precision": 0.9765584870417932,
      "eval_recall": 0.9857183257918553,
      "eval_runtime": 205.9645,
      "eval_samples_per_second": 71.653,
      "eval_steps_per_second": 2.243,
      "step": 14000
    },
    {
      "epoch": 7.745726495726496,
      "grad_norm": 0.0030930449720472097,
      "learning_rate": 4.4467338217338216e-05,
      "loss": 0.0003,
      "step": 14500
    },
    {
      "epoch": 7.745726495726496,
      "eval_f1": 0.9808375067597752,
      "eval_loss": 0.0013525269459933043,
      "eval_precision": 0.9785596997419658,
      "eval_recall": 0.9831259426847662,
      "eval_runtime": 122.7721,
      "eval_samples_per_second": 120.206,
      "eval_steps_per_second": 3.763,
      "step": 14500
    },
    {
      "epoch": 8.012820512820513,
      "grad_norm": 0.00297636934556067,
      "learning_rate": 4.427655677655678e-05,
      "loss": 0.0002,
      "step": 15000
    },
    {
      "epoch": 8.012820512820513,
      "eval_f1": 0.9820384687912502,
      "eval_loss": 0.0015595377190038562,
      "eval_precision": 0.9822236891738967,
      "eval_recall": 0.981853318250377,
      "eval_runtime": 208.3474,
      "eval_samples_per_second": 70.834,
      "eval_steps_per_second": 2.217,
      "step": 15000
    }
  ],
  "logging_steps": 500,
  "max_steps": 131040,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 70,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.2536694351387706e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}