trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.838709677419355,
  "eval_steps": 500,
  "global_step": 33,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8602150537634409,
      "grad_norm": 2.140043645872373,
      "learning_rate": 5e-06,
      "loss": 1.0259,
      "step": 10
    },
    {
      "epoch": 0.946236559139785,
      "eval_loss": 0.9281949996948242,
      "eval_runtime": 12.651,
      "eval_samples_per_second": 24.662,
      "eval_steps_per_second": 0.395,
      "step": 11
    },
    {
      "epoch": 1.7204301075268817,
      "grad_norm": 1.923788898589323,
      "learning_rate": 5e-06,
      "loss": 0.9242,
      "step": 20
    },
    {
      "epoch": 1.978494623655914,
      "eval_loss": 0.899213433265686,
      "eval_runtime": 12.5907,
      "eval_samples_per_second": 24.78,
      "eval_steps_per_second": 0.397,
      "step": 23
    },
    {
      "epoch": 2.5806451612903225,
      "grad_norm": 1.4418622622973827,
      "learning_rate": 5e-06,
      "loss": 0.8735,
      "step": 30
    },
    {
      "epoch": 2.838709677419355,
      "eval_loss": 0.8735513687133789,
      "eval_runtime": 12.203,
      "eval_samples_per_second": 25.567,
      "eval_steps_per_second": 0.41,
      "step": 33
    },
    {
      "epoch": 2.838709677419355,
      "step": 33,
      "total_flos": 55066849443840.0,
      "train_loss": 0.9298454053474196,
      "train_runtime": 2223.3969,
      "train_samples_per_second": 7.974,
      "train_steps_per_second": 0.015
    }
  ],
  "logging_steps": 10,
  "max_steps": 33,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 55066849443840.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}