{
  "best_global_step": 200,
  "best_metric": 0.9968940615653992,
  "best_model_checkpoint": "results_L0_12/R64_qwen_8b_structured/checkpoint-200",
  "epoch": 15.63768115942029,
  "eval_steps": 20,
  "global_step": 280,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5797101449275363,
      "grad_norm": 1.9165797233581543,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 3.2135,
      "step": 10
    },
    {
      "epoch": 1.1159420289855073,
      "grad_norm": 0.5033081769943237,
      "learning_rate": 3.518518518518519e-05,
      "loss": 2.5406,
      "step": 20
    },
    {
      "epoch": 1.1159420289855073,
      "eval_loss": 2.1649956703186035,
      "eval_runtime": 28.0718,
      "eval_samples_per_second": 19.735,
      "eval_steps_per_second": 2.494,
      "step": 20
    },
    {
      "epoch": 1.6956521739130435,
      "grad_norm": 0.7941814064979553,
      "learning_rate": 5.370370370370371e-05,
      "loss": 1.872,
      "step": 30
    },
    {
      "epoch": 2.2318840579710146,
      "grad_norm": 0.5307704210281372,
      "learning_rate": 7.222222222222222e-05,
      "loss": 1.6006,
      "step": 40
    },
    {
      "epoch": 2.2318840579710146,
      "eval_loss": 1.5740340948104858,
      "eval_runtime": 58.9447,
      "eval_samples_per_second": 9.399,
      "eval_steps_per_second": 1.188,
      "step": 40
    },
    {
      "epoch": 2.8115942028985508,
      "grad_norm": 0.18376459181308746,
      "learning_rate": 9.074074074074075e-05,
      "loss": 1.6336,
      "step": 50
    },
    {
      "epoch": 3.3478260869565215,
      "grad_norm": 0.24469375610351562,
      "learning_rate": 9.999797656920765e-05,
      "loss": 1.4624,
      "step": 60
    },
    {
      "epoch": 3.3478260869565215,
      "eval_loss": 1.4656774997711182,
      "eval_runtime": 59.8998,
      "eval_samples_per_second": 9.249,
      "eval_steps_per_second": 1.169,
      "step": 60
    },
    {
      "epoch": 3.927536231884058,
      "grad_norm": 0.29990342259407043,
      "learning_rate": 9.998179010548088e-05,
      "loss": 1.3827,
      "step": 70
    },
    {
      "epoch": 4.463768115942029,
      "grad_norm": 0.3905975818634033,
      "learning_rate": 9.994942241827156e-05,
      "loss": 1.3227,
      "step": 80
    },
    {
      "epoch": 4.463768115942029,
      "eval_loss": 1.403430700302124,
      "eval_runtime": 57.7438,
      "eval_samples_per_second": 9.594,
      "eval_steps_per_second": 1.212,
      "step": 80
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8633396029472351,
      "learning_rate": 9.990088398637167e-05,
      "loss": 1.311,
      "step": 90
    },
    {
      "epoch": 5.579710144927536,
      "grad_norm": 0.29788145422935486,
      "learning_rate": 9.983619052372848e-05,
      "loss": 1.3545,
      "step": 100
    },
    {
      "epoch": 5.579710144927536,
      "eval_loss": 1.367808222770691,
      "eval_runtime": 58.6669,
      "eval_samples_per_second": 9.443,
      "eval_steps_per_second": 1.193,
      "step": 100
    },
    {
      "epoch": 6.115942028985507,
      "grad_norm": 158.66375732421875,
      "learning_rate": 9.975536297435726e-05,
      "loss": 1.1799,
      "step": 110
    },
    {
      "epoch": 6.695652173913043,
      "grad_norm": 0.5346195101737976,
      "learning_rate": 9.965842750556091e-05,
      "loss": 1.163,
      "step": 120
    },
    {
      "epoch": 6.695652173913043,
      "eval_loss": 1.3441675901412964,
      "eval_runtime": 60.3552,
      "eval_samples_per_second": 9.179,
      "eval_steps_per_second": 1.16,
      "step": 120
    },
    {
      "epoch": 7.231884057971015,
      "grad_norm": 0.4752218425273895,
      "learning_rate": 9.95454154994584e-05,
      "loss": 1.152,
      "step": 130
    },
    {
      "epoch": 7.811594202898551,
      "grad_norm": 0.4084065556526184,
      "learning_rate": 9.941636354282511e-05,
      "loss": 1.209,
      "step": 140
    },
    {
      "epoch": 7.811594202898551,
      "eval_loss": 1.3283058404922485,
      "eval_runtime": 60.0993,
      "eval_samples_per_second": 9.218,
      "eval_steps_per_second": 1.165,
      "step": 140
    },
    {
      "epoch": 8.347826086956522,
      "grad_norm": 0.49419981241226196,
      "learning_rate": 9.927131341524809e-05,
      "loss": 1.0819,
      "step": 150
    },
    {
      "epoch": 8.927536231884059,
      "grad_norm": 0.5527717471122742,
      "learning_rate": 9.911031207560027e-05,
      "loss": 1.0363,
      "step": 160
    },
    {
      "epoch": 8.927536231884059,
      "eval_loss": 1.3205050230026245,
      "eval_runtime": 53.6751,
      "eval_samples_per_second": 10.321,
      "eval_steps_per_second": 1.304,
      "step": 160
    },
    {
      "epoch": 9.46376811594203,
      "grad_norm": 0.5900382399559021,
      "learning_rate": 9.893341164683788e-05,
      "loss": 1.0404,
      "step": 170
    },
    {
      "epoch": 10.0,
      "grad_norm": 1.0412408113479614,
      "learning_rate": 9.874066939912603e-05,
      "loss": 0.9826,
      "step": 180
    },
    {
      "epoch": 10.0,
      "eval_loss": 1.312466025352478,
      "eval_runtime": 60.1194,
      "eval_samples_per_second": 9.215,
      "eval_steps_per_second": 1.164,
      "step": 180
    },
    {
      "epoch": 10.579710144927537,
      "grad_norm": 0.5220209956169128,
      "learning_rate": 9.853214773129796e-05,
      "loss": 1.1468,
      "step": 190
    },
    {
      "epoch": 11.115942028985508,
      "grad_norm": 0.6074064373970032,
      "learning_rate": 9.830791415065382e-05,
      "loss": 1.0142,
      "step": 200
    },
    {
      "epoch": 11.115942028985508,
      "eval_loss": 0.9968940615653992,
      "eval_runtime": 65.3018,
      "eval_samples_per_second": 8.484,
      "eval_steps_per_second": 1.072,
      "step": 200
    },
    {
      "epoch": 11.695652173913043,
      "grad_norm": 0.6192728877067566,
      "learning_rate": 9.806804125110579e-05,
      "loss": 0.9176,
      "step": 210
    },
    {
      "epoch": 12.289855072463768,
      "grad_norm": 1.178308129310608,
      "learning_rate": 9.781260668967628e-05,
      "loss": 1.0747,
      "step": 220
    },
    {
      "epoch": 12.289855072463768,
      "eval_loss": 1.0378575325012207,
      "eval_runtime": 33.6087,
      "eval_samples_per_second": 16.484,
      "eval_steps_per_second": 2.083,
      "step": 220
    },
    {
      "epoch": 12.869565217391305,
      "grad_norm": 0.5823021531105042,
      "learning_rate": 9.754169316135712e-05,
      "loss": 0.9349,
      "step": 230
    },
    {
      "epoch": 13.405797101449275,
      "grad_norm": 1.0113617181777954,
      "learning_rate": 9.725538837233763e-05,
      "loss": 0.8412,
      "step": 240
    },
    {
      "epoch": 13.405797101449275,
      "eval_loss": 1.0512763261795044,
      "eval_runtime": 64.8372,
      "eval_samples_per_second": 8.544,
      "eval_steps_per_second": 1.08,
      "step": 240
    },
    {
      "epoch": 13.985507246376812,
      "grad_norm": 0.4507587254047394,
      "learning_rate": 9.695378501161045e-05,
      "loss": 0.8315,
      "step": 250
    },
    {
      "epoch": 14.521739130434783,
      "grad_norm": 0.5109806060791016,
      "learning_rate": 9.663698072096418e-05,
      "loss": 0.785,
      "step": 260
    },
    {
      "epoch": 14.521739130434783,
      "eval_loss": 1.081993579864502,
      "eval_runtime": 64.9546,
      "eval_samples_per_second": 8.529,
      "eval_steps_per_second": 1.078,
      "step": 260
    },
    {
      "epoch": 15.057971014492754,
      "grad_norm": 342.426025390625,
      "learning_rate": 9.630507806337259e-05,
      "loss": 0.7538,
      "step": 270
    },
    {
      "epoch": 15.63768115942029,
      "grad_norm": 0.9247072339057922,
      "learning_rate": 9.595818448979061e-05,
      "loss": 0.7405,
      "step": 280
    },
    {
      "epoch": 15.63768115942029,
      "eval_loss": 1.1149591207504272,
      "eval_runtime": 65.9467,
      "eval_samples_per_second": 8.401,
      "eval_steps_per_second": 1.061,
      "step": 280
    }
  ],
  "logging_steps": 10,
  "max_steps": 1800,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 20,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 4,
        "early_stopping_threshold": 0.01
      },
      "attributes": {
        "early_stopping_patience_counter": 4
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8.238044479018107e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}