{
  "best_metric": 0.037236765027046204,
  "best_model_checkpoint": "doc-topic-model_eval-02_train-00/checkpoint-14000",
  "epoch": 9.368836291913215,
  "eval_steps": 1000,
  "global_step": 19000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2465483234714004,
      "grad_norm": 0.32839319109916687,
      "learning_rate": 1.9950690335305722e-05,
      "loss": 0.1668,
      "step": 500
    },
    {
      "epoch": 0.4930966469428008,
      "grad_norm": 0.4088590443134308,
      "learning_rate": 1.9901380670611442e-05,
      "loss": 0.0929,
      "step": 1000
    },
    {
      "epoch": 0.4930966469428008,
      "eval_accuracy": 0.9813654208653996,
      "eval_f1": 0.0,
      "eval_loss": 0.09118108451366425,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 12.1097,
      "eval_samples_per_second": 669.709,
      "eval_steps_per_second": 2.642,
      "step": 1000
    },
    {
      "epoch": 0.7396449704142012,
      "grad_norm": 0.33391374349594116,
      "learning_rate": 1.9852071005917162e-05,
      "loss": 0.0874,
      "step": 1500
    },
    {
      "epoch": 0.9861932938856016,
      "grad_norm": 0.3162483870983124,
      "learning_rate": 1.980276134122288e-05,
      "loss": 0.0785,
      "step": 2000
    },
    {
      "epoch": 0.9861932938856016,
      "eval_accuracy": 0.9813654208653996,
      "eval_f1": 0.0,
      "eval_loss": 0.07077465206384659,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 12.1773,
      "eval_samples_per_second": 665.993,
      "eval_steps_per_second": 2.628,
      "step": 2000
    },
    {
      "epoch": 1.232741617357002,
      "grad_norm": 0.31390729546546936,
      "learning_rate": 1.9753451676528602e-05,
      "loss": 0.0684,
      "step": 2500
    },
    {
      "epoch": 1.4792899408284024,
      "grad_norm": 0.40165144205093384,
      "learning_rate": 1.9704142011834322e-05,
      "loss": 0.0622,
      "step": 3000
    },
    {
      "epoch": 1.4792899408284024,
      "eval_accuracy": 0.9822994174922404,
      "eval_f1": 0.11091336228376165,
      "eval_loss": 0.05758311599493027,
      "eval_precision": 0.8665183537263627,
      "eval_recall": 0.059248554913294796,
      "eval_runtime": 14.3679,
      "eval_samples_per_second": 564.454,
      "eval_steps_per_second": 2.227,
      "step": 3000
    },
    {
      "epoch": 1.725838264299803,
      "grad_norm": 0.4135937988758087,
      "learning_rate": 1.965483234714004e-05,
      "loss": 0.0584,
      "step": 3500
    },
    {
      "epoch": 1.972386587771203,
      "grad_norm": 0.324859082698822,
      "learning_rate": 1.9605522682445763e-05,
      "loss": 0.0542,
      "step": 4000
    },
    {
      "epoch": 1.972386587771203,
      "eval_accuracy": 0.9842311889677849,
      "eval_f1": 0.3405642484589853,
      "eval_loss": 0.05005618929862976,
      "eval_precision": 0.7714822771213748,
      "eval_recall": 0.21851232126559172,
      "eval_runtime": 14.3627,
      "eval_samples_per_second": 564.655,
      "eval_steps_per_second": 2.228,
      "step": 4000
    },
    {
      "epoch": 2.2189349112426036,
      "grad_norm": 0.3279891312122345,
      "learning_rate": 1.9556213017751483e-05,
      "loss": 0.0488,
      "step": 4500
    },
    {
      "epoch": 2.465483234714004,
      "grad_norm": 0.4424910843372345,
      "learning_rate": 1.95069033530572e-05,
      "loss": 0.048,
      "step": 5000
    },
    {
      "epoch": 2.465483234714004,
      "eval_accuracy": 0.9852799863939793,
      "eval_f1": 0.42770553229005953,
      "eval_loss": 0.046073004603385925,
      "eval_precision": 0.7762,
      "eval_recall": 0.29517797383632494,
      "eval_runtime": 14.498,
      "eval_samples_per_second": 559.388,
      "eval_steps_per_second": 2.207,
      "step": 5000
    },
    {
      "epoch": 2.712031558185404,
      "grad_norm": 0.3125520646572113,
      "learning_rate": 1.9457593688362923e-05,
      "loss": 0.0453,
      "step": 5500
    },
    {
      "epoch": 2.9585798816568047,
      "grad_norm": 0.4686921536922455,
      "learning_rate": 1.940828402366864e-05,
      "loss": 0.0436,
      "step": 6000
    },
    {
      "epoch": 2.9585798816568047,
      "eval_accuracy": 0.9861473702113185,
      "eval_f1": 0.5112022404480896,
      "eval_loss": 0.04339677840471268,
      "eval_precision": 0.7463492990654206,
      "eval_recall": 0.38872832369942195,
      "eval_runtime": 14.5997,
      "eval_samples_per_second": 555.492,
      "eval_steps_per_second": 2.192,
      "step": 6000
    },
    {
      "epoch": 3.2051282051282053,
      "grad_norm": 0.3562796711921692,
      "learning_rate": 1.935897435897436e-05,
      "loss": 0.0412,
      "step": 6500
    },
    {
      "epoch": 3.4516765285996054,
      "grad_norm": 0.24857360124588013,
      "learning_rate": 1.930966469428008e-05,
      "loss": 0.0384,
      "step": 7000
    },
    {
      "epoch": 3.4516765285996054,
      "eval_accuracy": 0.9866873591564267,
      "eval_f1": 0.5495612142137822,
      "eval_loss": 0.0414077453315258,
      "eval_precision": 0.7436729396495781,
      "eval_recall": 0.4358077274110131,
      "eval_runtime": 14.527,
      "eval_samples_per_second": 558.269,
      "eval_steps_per_second": 2.203,
      "step": 7000
    },
    {
      "epoch": 3.698224852071006,
      "grad_norm": 0.47940972447395325,
      "learning_rate": 1.92603550295858e-05,
      "loss": 0.0394,
      "step": 7500
    },
    {
      "epoch": 3.9447731755424065,
      "grad_norm": 0.39106041193008423,
      "learning_rate": 1.921104536489152e-05,
      "loss": 0.0385,
      "step": 8000
    },
    {
      "epoch": 3.9447731755424065,
      "eval_accuracy": 0.9866717689244158,
      "eval_f1": 0.5362919132149901,
      "eval_loss": 0.040173791348934174,
      "eval_precision": 0.7624789680314077,
      "eval_recall": 0.41359902646790386,
      "eval_runtime": 14.5683,
      "eval_samples_per_second": 556.687,
      "eval_steps_per_second": 2.197,
      "step": 8000
    },
    {
      "epoch": 4.191321499013807,
      "grad_norm": 0.4470931887626648,
      "learning_rate": 1.916173570019724e-05,
      "loss": 0.0342,
      "step": 8500
    },
    {
      "epoch": 4.437869822485207,
      "grad_norm": 0.33709824085235596,
      "learning_rate": 1.911242603550296e-05,
      "loss": 0.0343,
      "step": 9000
    },
    {
      "epoch": 4.437869822485207,
      "eval_accuracy": 0.9869977465028275,
      "eval_f1": 0.5633092155369384,
      "eval_loss": 0.03959447145462036,
      "eval_precision": 0.7527989821882952,
      "eval_recall": 0.45003042287800427,
      "eval_runtime": 14.5424,
      "eval_samples_per_second": 557.678,
      "eval_steps_per_second": 2.2,
      "step": 9000
    },
    {
      "epoch": 4.684418145956608,
      "grad_norm": 0.3310143053531647,
      "learning_rate": 1.906311637080868e-05,
      "loss": 0.0343,
      "step": 9500
    },
    {
      "epoch": 4.930966469428008,
      "grad_norm": 0.5156289935112,
      "learning_rate": 1.90138067061144e-05,
      "loss": 0.0343,
      "step": 10000
    },
    {
      "epoch": 4.930966469428008,
      "eval_accuracy": 0.9872230962200774,
      "eval_f1": 0.57721708952774,
      "eval_loss": 0.03884171321988106,
      "eval_precision": 0.752782874617737,
      "eval_recall": 0.46805597809552785,
      "eval_runtime": 14.5796,
      "eval_samples_per_second": 556.255,
      "eval_steps_per_second": 2.195,
      "step": 10000
    },
    {
      "epoch": 5.177514792899408,
      "grad_norm": 0.3030126988887787,
      "learning_rate": 1.896459566074951e-05,
      "loss": 0.0314,
      "step": 10500
    },
    {
      "epoch": 5.424063116370808,
      "grad_norm": 0.49180835485458374,
      "learning_rate": 1.891528599605523e-05,
      "loss": 0.0304,
      "step": 11000
    },
    {
      "epoch": 5.424063116370808,
      "eval_accuracy": 0.9870983743639894,
      "eval_f1": 0.5816059199338144,
      "eval_loss": 0.038809314370155334,
      "eval_precision": 0.7349285631316065,
      "eval_recall": 0.48121387283236994,
      "eval_runtime": 14.3105,
      "eval_samples_per_second": 566.716,
      "eval_steps_per_second": 2.236,
      "step": 11000
    },
    {
      "epoch": 5.670611439842209,
      "grad_norm": 0.4685238003730774,
      "learning_rate": 1.886597633136095e-05,
      "loss": 0.0297,
      "step": 11500
    },
    {
      "epoch": 5.9171597633136095,
      "grad_norm": 0.46912193298339844,
      "learning_rate": 1.881666666666667e-05,
      "loss": 0.0299,
      "step": 12000
    },
    {
      "epoch": 5.9171597633136095,
      "eval_accuracy": 0.987516476040648,
      "eval_f1": 0.6071364852809991,
      "eval_loss": 0.03742339462041855,
      "eval_precision": 0.7340379637618637,
      "eval_recall": 0.5176452692424703,
      "eval_runtime": 14.2866,
      "eval_samples_per_second": 567.665,
      "eval_steps_per_second": 2.24,
      "step": 12000
    },
    {
      "epoch": 6.16370808678501,
      "grad_norm": 0.3353117108345032,
      "learning_rate": 1.876735700197239e-05,
      "loss": 0.0276,
      "step": 12500
    },
    {
      "epoch": 6.410256410256411,
      "grad_norm": 0.2552751898765564,
      "learning_rate": 1.8718145956607497e-05,
      "loss": 0.0265,
      "step": 13000
    },
    {
      "epoch": 6.410256410256411,
      "eval_accuracy": 0.9874682880507958,
      "eval_f1": 0.6134813778632628,
      "eval_loss": 0.03768361359834671,
      "eval_precision": 0.7213199013157895,
      "eval_recall": 0.533693337389717,
      "eval_runtime": 14.3396,
      "eval_samples_per_second": 565.567,
      "eval_steps_per_second": 2.232,
      "step": 13000
    },
    {
      "epoch": 6.65680473372781,
      "grad_norm": 0.2560911774635315,
      "learning_rate": 1.8668836291913217e-05,
      "loss": 0.0273,
      "step": 13500
    },
    {
      "epoch": 6.903353057199211,
      "grad_norm": 0.4542127549648285,
      "learning_rate": 1.8619526627218937e-05,
      "loss": 0.0261,
      "step": 14000
    },
    {
      "epoch": 6.903353057199211,
      "eval_accuracy": 0.9876454497781935,
      "eval_f1": 0.6116630284670558,
      "eval_loss": 0.037236765027046204,
      "eval_precision": 0.7382514248843962,
      "eval_recall": 0.5221326437480985,
      "eval_runtime": 14.3666,
      "eval_samples_per_second": 564.504,
      "eval_steps_per_second": 2.227,
      "step": 14000
    },
    {
      "epoch": 7.149901380670611,
      "grad_norm": 0.2716215252876282,
      "learning_rate": 1.8570216962524657e-05,
      "loss": 0.025,
      "step": 14500
    },
    {
      "epoch": 7.396449704142012,
      "grad_norm": 0.2959195077419281,
      "learning_rate": 1.8520907297830377e-05,
      "loss": 0.0236,
      "step": 15000
    },
    {
      "epoch": 7.396449704142012,
      "eval_accuracy": 0.9876511189534702,
      "eval_f1": 0.6207286814956688,
      "eval_loss": 0.037736888974905014,
      "eval_precision": 0.7256997455470738,
      "eval_recall": 0.5422878004259203,
      "eval_runtime": 14.6872,
      "eval_samples_per_second": 552.183,
      "eval_steps_per_second": 2.179,
      "step": 15000
    },
    {
      "epoch": 7.642998027613412,
      "grad_norm": 0.44261014461517334,
      "learning_rate": 1.8471696252465485e-05,
      "loss": 0.0225,
      "step": 15500
    },
    {
      "epoch": 7.889546351084813,
      "grad_norm": 0.721012532711029,
      "learning_rate": 1.8422386587771205e-05,
      "loss": 0.0236,
      "step": 16000
    },
    {
      "epoch": 7.889546351084813,
      "eval_accuracy": 0.9878353671499639,
      "eval_f1": 0.6227750186788555,
      "eval_loss": 0.03772629797458649,
      "eval_precision": 0.7376366475793857,
      "eval_recall": 0.5388652266504411,
      "eval_runtime": 14.358,
      "eval_samples_per_second": 564.84,
      "eval_steps_per_second": 2.229,
      "step": 16000
    },
    {
      "epoch": 8.136094674556213,
      "grad_norm": 0.2590983510017395,
      "learning_rate": 1.8373076923076926e-05,
      "loss": 0.0219,
      "step": 16500
    },
    {
      "epoch": 8.382642998027613,
      "grad_norm": 0.6246200203895569,
      "learning_rate": 1.8323767258382646e-05,
      "loss": 0.0215,
      "step": 17000
    },
    {
      "epoch": 8.382642998027613,
      "eval_accuracy": 0.9878552092634324,
      "eval_f1": 0.6335685268334402,
      "eval_loss": 0.03787342831492424,
      "eval_precision": 0.7236495066914135,
      "eval_recall": 0.5634317006388805,
      "eval_runtime": 14.4229,
      "eval_samples_per_second": 562.299,
      "eval_steps_per_second": 2.219,
      "step": 17000
    },
    {
      "epoch": 8.629191321499015,
      "grad_norm": 0.46473428606987,
      "learning_rate": 1.8274457593688362e-05,
      "loss": 0.0204,
      "step": 17500
    },
    {
      "epoch": 8.875739644970414,
      "grad_norm": 0.5553109049797058,
      "learning_rate": 1.822524654832347e-05,
      "loss": 0.0216,
      "step": 18000
    },
    {
      "epoch": 8.875739644970414,
      "eval_accuracy": 0.987812690448857,
      "eval_f1": 0.6330075540950023,
      "eval_loss": 0.03817650303244591,
      "eval_precision": 0.7211903141106681,
      "eval_recall": 0.5640401581989656,
      "eval_runtime": 14.4328,
      "eval_samples_per_second": 561.914,
      "eval_steps_per_second": 2.217,
      "step": 18000
    },
    {
      "epoch": 9.122287968441814,
      "grad_norm": 0.3819935917854309,
      "learning_rate": 1.8175936883629194e-05,
      "loss": 0.0194,
      "step": 18500
    },
    {
      "epoch": 9.368836291913215,
      "grad_norm": 0.5073143243789673,
      "learning_rate": 1.8126627218934914e-05,
      "loss": 0.0177,
      "step": 19000
    },
    {
      "epoch": 9.368836291913215,
      "eval_accuracy": 0.9878807205521777,
      "eval_f1": 0.6370389235536313,
      "eval_loss": 0.038153521716594696,
      "eval_precision": 0.7207761021995965,
      "eval_recall": 0.5707331913599026,
      "eval_runtime": 14.4357,
      "eval_samples_per_second": 561.802,
      "eval_steps_per_second": 2.217,
      "step": 19000
    }
  ],
  "logging_steps": 500,
  "max_steps": 202800,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 1000,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 430075849523328.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}