{
  "best_global_step": 20000,
  "best_metric": 1.0081566159669819e-07,
  "best_model_checkpoint": "check-point-qwen2.5/checkpoint-20000",
  "epoch": 4.793911428057769,
  "eval_steps": 2000,
  "global_step": 20000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.479415113561455,
      "eval_entropy": 4.887580871582031e-05,
      "eval_loss": 3.042736807401525e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 32768000.0,
      "eval_runtime": 24.1477,
      "eval_samples_per_second": 42.737,
      "eval_steps_per_second": 10.684,
      "step": 2000
    },
    {
      "epoch": 0.95883022712291,
      "eval_entropy": 2.6106834411621094e-05,
      "eval_loss": 1.529578639747342e-06,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 65536000.0,
      "eval_runtime": 24.154,
      "eval_samples_per_second": 42.726,
      "eval_steps_per_second": 10.681,
      "step": 4000
    },
    {
      "entropy": 0.0053923043380598315,
      "epoch": 1.0,
      "grad_norm": 3.886222839355469e-05,
      "learning_rate": 0.00016078034682080926,
      "loss": 0.0052,
      "mean_token_accuracy": 0.9988916747514459,
      "num_tokens": 68347904.0,
      "step": 4172
    },
    {
      "epoch": 1.43818541379517,
      "eval_entropy": 1.436471939086914e-05,
      "eval_loss": 7.912742603366496e-07,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 98297856.0,
      "eval_runtime": 25.3848,
      "eval_samples_per_second": 40.654,
      "eval_steps_per_second": 10.164,
      "step": 6000
    },
    {
      "epoch": 1.9176005273566248,
      "eval_entropy": 8.404254913330078e-06,
      "eval_loss": 4.2584713355608983e-07,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 131065856.0,
      "eval_runtime": 24.1529,
      "eval_samples_per_second": 42.728,
      "eval_steps_per_second": 10.682,
      "step": 8000
    },
    {
      "entropy": 1.4303378792515465e-05,
      "epoch": 2.0,
      "grad_norm": 9.894371032714844e-06,
      "learning_rate": 0.00012058766859344895,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 136695808.0,
      "step": 8344
    },
    {
      "epoch": 2.3969557140288846,
      "eval_entropy": 5.3942203521728516e-06,
      "eval_loss": 2.4808102239148866e-07,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 163827712.0,
      "eval_runtime": 24.1649,
      "eval_samples_per_second": 42.707,
      "eval_steps_per_second": 10.677,
      "step": 10000
    },
    {
      "epoch": 2.87637082759034,
      "eval_entropy": 3.904104232788086e-06,
      "eval_loss": 1.6367992827781563e-07,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 196595712.0,
      "eval_runtime": 25.4048,
      "eval_samples_per_second": 40.622,
      "eval_steps_per_second": 10.156,
      "step": 12000
    },
    {
      "entropy": 5.196735094097524e-06,
      "epoch": 3.0,
      "grad_norm": 3.7103891372680664e-06,
      "learning_rate": 8.039499036608865e-05,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 205043712.0,
      "step": 12516
    },
    {
      "epoch": 3.3557260142625998,
      "eval_entropy": 3.248453140258789e-06,
      "eval_loss": 1.2468079546579247e-07,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 229357568.0,
      "eval_runtime": 25.3871,
      "eval_samples_per_second": 40.65,
      "eval_steps_per_second": 10.163,
      "step": 14000
    },
    {
      "epoch": 3.8351411278240546,
      "eval_entropy": 3.0547380447387695e-06,
      "eval_loss": 1.0977963427194481e-07,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 262125568.0,
      "eval_runtime": 24.1299,
      "eval_samples_per_second": 42.769,
      "eval_steps_per_second": 10.692,
      "step": 16000
    },
    {
      "entropy": 3.2122802968693013e-06,
      "epoch": 4.0,
      "grad_norm": 3.0994415283203125e-06,
      "learning_rate": 4.020231213872833e-05,
      "loss": 0.0,
      "mean_token_accuracy": 1.0,
      "num_tokens": 273391616.0,
      "step": 16688
    },
    {
      "epoch": 4.314496314496314,
      "eval_entropy": 2.9355287551879883e-06,
      "eval_loss": 1.0197981481496754e-07,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 294887424.0,
      "eval_runtime": 24.1367,
      "eval_samples_per_second": 42.756,
      "eval_steps_per_second": 10.689,
      "step": 18000
    },
    {
      "epoch": 4.793911428057769,
      "eval_entropy": 2.9206275939941406e-06,
      "eval_loss": 1.0081566159669819e-07,
      "eval_mean_token_accuracy": 1.0,
      "eval_num_tokens": 327655424.0,
      "eval_runtime": 24.1342,
      "eval_samples_per_second": 42.761,
      "eval_steps_per_second": 10.69,
      "step": 20000
    }
  ],
  "logging_steps": 4172,
  "max_steps": 20860,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 2000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.576041244991947e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}