{
  "best_global_step": 600,
  "best_metric": 2.3092410564422607,
  "best_model_checkpoint": "gpt-oss-20b-depression-detection/checkpoint-600",
  "epoch": 1.5969353764157228,
  "eval_steps": 100,
  "global_step": 600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 3.24503686606884,
      "epoch": 0.06662225183211193,
      "grad_norm": 2.454188108444214,
      "learning_rate": 0.00031578947368421053,
      "loss": 3.3772,
      "mean_token_accuracy": 0.3810029237344861,
      "num_tokens": 35980.0,
      "step": 25
    },
    {
      "entropy": 2.3023448407649996,
      "epoch": 0.13324450366422386,
      "grad_norm": 1.2468138933181763,
      "learning_rate": 0.0004997072385268466,
      "loss": 2.5437,
      "mean_token_accuracy": 0.5166339360177516,
      "num_tokens": 80261.0,
      "step": 50
    },
    {
      "entropy": 2.2647578102350234,
      "epoch": 0.19986675549633579,
      "grad_norm": 1.1340042352676392,
      "learning_rate": 0.0004968702467935468,
      "loss": 2.5339,
      "mean_token_accuracy": 0.5262378923594951,
      "num_tokens": 120732.0,
      "step": 75
    },
    {
      "entropy": 2.384706051647663,
      "epoch": 0.2664890073284477,
      "grad_norm": 1.2488991022109985,
      "learning_rate": 0.0004910491498244077,
      "loss": 2.5944,
      "mean_token_accuracy": 0.502809617370367,
      "num_tokens": 164621.0,
      "step": 100
    },
    {
      "epoch": 0.2664890073284477,
      "eval_entropy": 2.310890290975571,
      "eval_loss": 2.350224256515503,
      "eval_mean_token_accuracy": 0.5086323776245117,
      "eval_num_tokens": 164621.0,
      "eval_runtime": 198.3169,
      "eval_samples_per_second": 5.042,
      "eval_steps_per_second": 5.042,
      "step": 100
    },
    {
      "entropy": 2.2780907002091406,
      "epoch": 0.3331112591605596,
      "grad_norm": 1.2890889644622803,
      "learning_rate": 0.0004823143115704097,
      "loss": 2.4955,
      "mean_token_accuracy": 0.5228405058383941,
      "num_tokens": 202317.0,
      "step": 125
    },
    {
      "entropy": 2.252679335474968,
      "epoch": 0.39973351099267157,
      "grad_norm": 1.3215383291244507,
      "learning_rate": 0.0004707713165517877,
      "loss": 2.5347,
      "mean_token_accuracy": 0.5338048176467418,
      "num_tokens": 242702.0,
      "step": 150
    },
    {
      "entropy": 2.2355666947364807,
      "epoch": 0.46635576282478347,
      "grad_norm": 1.1142017841339111,
      "learning_rate": 0.00045655969357899875,
      "loss": 2.5003,
      "mean_token_accuracy": 0.5333976492285728,
      "num_tokens": 278791.0,
      "step": 175
    },
    {
      "entropy": 2.2577033042907715,
      "epoch": 0.5329780146568954,
      "grad_norm": 0.8657254576683044,
      "learning_rate": 0.0004398512291636768,
      "loss": 2.5257,
      "mean_token_accuracy": 0.5266946247220039,
      "num_tokens": 318029.0,
      "step": 200
    },
    {
      "epoch": 0.5329780146568954,
      "eval_entropy": 2.2311146975159644,
      "eval_loss": 2.343594789505005,
      "eval_mean_token_accuracy": 0.5092586134076118,
      "eval_num_tokens": 318029.0,
      "eval_runtime": 198.601,
      "eval_samples_per_second": 5.035,
      "eval_steps_per_second": 5.035,
      "step": 200
    },
    {
      "entropy": 2.3073052656650543,
      "epoch": 0.5996002664890073,
      "grad_norm": 1.1857596635818481,
      "learning_rate": 0.00042084789100663714,
      "loss": 2.5791,
      "mean_token_accuracy": 0.5132052838802338,
      "num_tokens": 358254.0,
      "step": 225
    },
    {
      "entropy": 2.3000676898658274,
      "epoch": 0.6662225183211192,
      "grad_norm": 1.1998451948165894,
      "learning_rate": 0.0003997793866632835,
      "loss": 2.5177,
      "mean_token_accuracy": 0.5144143861532211,
      "num_tokens": 400981.0,
      "step": 250
    },
    {
      "entropy": 2.27166977584362,
      "epoch": 0.7328447701532311,
      "grad_norm": 0.8986361026763916,
      "learning_rate": 0.000376900386896654,
      "loss": 2.5648,
      "mean_token_accuracy": 0.5221457535028458,
      "num_tokens": 442783.0,
      "step": 275
    },
    {
      "entropy": 2.3170408844947814,
      "epoch": 0.7994670219853431,
      "grad_norm": 1.1080683469772339,
      "learning_rate": 0.00035248744728151344,
      "loss": 2.578,
      "mean_token_accuracy": 0.5194095669686795,
      "num_tokens": 482598.0,
      "step": 300
    },
    {
      "epoch": 0.7994670219853431,
      "eval_entropy": 2.3874710043668745,
      "eval_loss": 2.3318331241607666,
      "eval_mean_token_accuracy": 0.5110126642882824,
      "eval_num_tokens": 482598.0,
      "eval_runtime": 198.6133,
      "eval_samples_per_second": 5.035,
      "eval_steps_per_second": 5.035,
      "step": 300
    },
    {
      "entropy": 2.291095963716507,
      "epoch": 0.866089273817455,
      "grad_norm": 1.20793879032135,
      "learning_rate": 0.00032683566527036574,
      "loss": 2.5405,
      "mean_token_accuracy": 0.5264704664051533,
      "num_tokens": 523311.0,
      "step": 325
    },
    {
      "entropy": 2.3025157406926153,
      "epoch": 0.9327115256495669,
      "grad_norm": 1.2968449592590332,
      "learning_rate": 0.00030025511312992733,
      "loss": 2.5486,
      "mean_token_accuracy": 0.519840395450592,
      "num_tokens": 564020.0,
      "step": 350
    },
    {
      "entropy": 2.20314584761858,
      "epoch": 0.9993337774816788,
      "grad_norm": 1.0814260244369507,
      "learning_rate": 0.0002730670898658255,
      "loss": 2.4991,
      "mean_token_accuracy": 0.5384732019901276,
      "num_tokens": 603285.0,
      "step": 375
    },
    {
      "entropy": 2.14934338091575,
      "epoch": 1.0639573617588274,
      "grad_norm": 1.1806567907333374,
      "learning_rate": 0.00024560023744131163,
      "loss": 2.3871,
      "mean_token_accuracy": 0.5480183097197837,
      "num_tokens": 641871.0,
      "step": 400
    },
    {
      "epoch": 1.0639573617588274,
      "eval_entropy": 2.133557476878166,
      "eval_loss": 2.323554277420044,
      "eval_mean_token_accuracy": 0.5134236664175987,
      "eval_num_tokens": 641871.0,
      "eval_runtime": 198.3508,
      "eval_samples_per_second": 5.042,
      "eval_steps_per_second": 5.042,
      "step": 400
    },
    {
      "entropy": 2.1422861018776893,
      "epoch": 1.1305796135909394,
      "grad_norm": 1.1475857496261597,
      "learning_rate": 0.0002181865682361542,
      "loss": 2.3492,
      "mean_token_accuracy": 0.5404227763414383,
      "num_tokens": 684830.0,
      "step": 425
    },
    {
      "entropy": 2.206546678543091,
      "epoch": 1.1972018654230512,
      "grad_norm": 1.206028938293457,
      "learning_rate": 0.00019115745176478944,
      "loss": 2.3688,
      "mean_token_accuracy": 0.5313847596943378,
      "num_tokens": 723212.0,
      "step": 450
    },
    {
      "entropy": 2.1671278363466264,
      "epoch": 1.2638241172551632,
      "grad_norm": 1.1863069534301758,
      "learning_rate": 0.00016483960916526414,
      "loss": 2.3838,
      "mean_token_accuracy": 0.5400173373520374,
      "num_tokens": 768577.0,
      "step": 475
    },
    {
      "entropy": 2.1349931293725968,
      "epoch": 1.3304463690872752,
      "grad_norm": 0.9461486339569092,
      "learning_rate": 0.0001395511638765804,
      "loss": 2.323,
      "mean_token_accuracy": 0.5418752540647983,
      "num_tokens": 809245.0,
      "step": 500
    },
    {
      "epoch": 1.3304463690872752,
      "eval_entropy": 2.243591215968132,
      "eval_loss": 2.3116395473480225,
      "eval_mean_token_accuracy": 0.5138829910159111,
      "eval_num_tokens": 809245.0,
      "eval_runtime": 198.4593,
      "eval_samples_per_second": 5.039,
      "eval_steps_per_second": 5.039,
      "step": 500
    },
    {
      "entropy": 2.086217446625233,
      "epoch": 1.397068620919387,
      "grad_norm": 1.2557684183120728,
      "learning_rate": 0.00011559779624285562,
      "loss": 2.351,
      "mean_token_accuracy": 0.5539429695904254,
      "num_tokens": 848203.0,
      "step": 525
    },
    {
      "entropy": 2.1425654873251916,
      "epoch": 1.463690872751499,
      "grad_norm": 1.4949860572814941,
      "learning_rate": 9.326904852647344e-05,
      "loss": 2.3622,
      "mean_token_accuracy": 0.5448546643555164,
      "num_tokens": 888372.0,
      "step": 550
    },
    {
      "entropy": 2.138375435769558,
      "epoch": 1.530313124583611,
      "grad_norm": 1.079410195350647,
      "learning_rate": 7.283482499429295e-05,
      "loss": 2.361,
      "mean_token_accuracy": 0.5505425909161568,
      "num_tokens": 928674.0,
      "step": 575
    },
    {
      "entropy": 2.126588108241558,
      "epoch": 1.5969353764157228,
      "grad_norm": 1.149130940437317,
      "learning_rate": 5.454212938299255e-05,
      "loss": 2.3546,
      "mean_token_accuracy": 0.5477004034817219,
      "num_tokens": 968761.0,
      "step": 600
    },
    {
      "epoch": 1.5969353764157228,
      "eval_entropy": 2.201212148785591,
      "eval_loss": 2.3092410564422607,
      "eval_mean_token_accuracy": 0.5150551037490367,
      "eval_num_tokens": 968761.0,
      "eval_runtime": 198.5206,
      "eval_samples_per_second": 5.037,
      "eval_steps_per_second": 5.037,
      "step": 600
    }
  ],
  "logging_steps": 25,
  "max_steps": 752,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1829472171823347e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}