trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.588235294117647,
  "eval_steps": 500,
  "global_step": 32,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5882352941176471,
      "grad_norm": 0.7869846357918158,
      "learning_rate": 4.8214285714285716e-05,
      "loss": 1.1594,
      "mean_token_accuracy": 0.7533773094415664,
      "num_tokens": 2505918.0,
      "step": 5
    },
    {
      "epoch": 1.1176470588235294,
      "grad_norm": 0.7283705538721326,
      "learning_rate": 3.928571428571429e-05,
      "loss": 0.9727,
      "mean_token_accuracy": 0.7840807305441962,
      "num_tokens": 4791423.0,
      "step": 10
    },
    {
      "epoch": 1.7058823529411766,
      "grad_norm": 0.29849651744475264,
      "learning_rate": 3.0357142857142857e-05,
      "loss": 0.8064,
      "mean_token_accuracy": 0.8134194314479828,
      "num_tokens": 7330272.0,
      "step": 15
    },
    {
      "epoch": 2.235294117647059,
      "grad_norm": 0.2833815146754528,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 0.7888,
      "mean_token_accuracy": 0.8178626464472877,
      "num_tokens": 9618952.0,
      "step": 20
    },
    {
      "epoch": 2.8235294117647056,
      "grad_norm": 0.24905338345002254,
      "learning_rate": 1.25e-05,
      "loss": 0.6827,
      "mean_token_accuracy": 0.8375016838312149,
      "num_tokens": 12114377.0,
      "step": 25
    },
    {
      "epoch": 3.3529411764705883,
      "grad_norm": 0.2945591740712261,
      "learning_rate": 3.5714285714285714e-06,
      "loss": 0.6889,
      "mean_token_accuracy": 0.8388072550296783,
      "num_tokens": 14409009.0,
      "step": 30
    },
    {
      "epoch": 3.588235294117647,
      "mean_token_accuracy": 0.8461483642458916,
      "num_tokens": 15425424.0,
      "step": 32,
      "total_flos": 22955630264320.0,
      "train_loss": 0.8402301967144012,
      "train_runtime": 202.7562,
      "train_samples_per_second": 10.633,
      "train_steps_per_second": 0.158
    }
  ],
  "logging_steps": 5,
  "max_steps": 32,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 22955630264320.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}