trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 6.25,
  "eval_steps": 25,
  "global_step": 75,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 1.6361405849456787,
      "learning_rate": 0.00045454545454545455,
      "loss": 1.518,
      "step": 10
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 1.2719820737838745,
      "learning_rate": 0.00040404040404040404,
      "loss": 0.719,
      "step": 20
    },
    {
      "epoch": 2.0833333333333335,
      "eval_loss": 0.9498233795166016,
      "eval_runtime": 2.6397,
      "eval_samples_per_second": 4.546,
      "eval_steps_per_second": 0.758,
      "step": 25
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.8773032426834106,
      "learning_rate": 0.00035353535353535354,
      "loss": 0.5112,
      "step": 30
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 0.5878697037696838,
      "learning_rate": 0.00030303030303030303,
      "loss": 0.376,
      "step": 40
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 0.5939679741859436,
      "learning_rate": 0.0002525252525252525,
      "loss": 0.2813,
      "step": 50
    },
    {
      "epoch": 4.166666666666667,
      "eval_loss": 1.3289717435836792,
      "eval_runtime": 2.6487,
      "eval_samples_per_second": 4.53,
      "eval_steps_per_second": 0.755,
      "step": 50
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.341347873210907,
      "learning_rate": 0.00020202020202020202,
      "loss": 0.2676,
      "step": 60
    },
    {
      "epoch": 5.833333333333333,
      "grad_norm": 0.447550505399704,
      "learning_rate": 0.00015151515151515152,
      "loss": 0.2165,
      "step": 70
    },
    {
      "epoch": 6.25,
      "eval_loss": 1.4123486280441284,
      "eval_runtime": 2.6494,
      "eval_samples_per_second": 4.529,
      "eval_steps_per_second": 0.755,
      "step": 75
    }
  ],
  "logging_steps": 10,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 25,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 480396498591744.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}