trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2857142857142857,
      "grad_norm": 10.168943192840892,
      "learning_rate": 5e-06,
      "loss": 0.6911,
      "step": 10
    },
    {
      "epoch": 0.5714285714285714,
      "grad_norm": 1.3016207580367245,
      "learning_rate": 5e-06,
      "loss": 0.6052,
      "step": 20
    },
    {
      "epoch": 0.8571428571428571,
      "grad_norm": 1.4046209891317623,
      "learning_rate": 5e-06,
      "loss": 0.559,
      "step": 30
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.5510295033454895,
      "eval_runtime": 3.8699,
      "eval_samples_per_second": 237.99,
      "eval_steps_per_second": 1.034,
      "step": 35
    },
    {
      "epoch": 1.1428571428571428,
      "grad_norm": 2.5391209192392252,
      "learning_rate": 5e-06,
      "loss": 0.5302,
      "step": 40
    },
    {
      "epoch": 1.4285714285714286,
      "grad_norm": 0.9905615818164379,
      "learning_rate": 5e-06,
      "loss": 0.5087,
      "step": 50
    },
    {
      "epoch": 1.7142857142857144,
      "grad_norm": 0.8530161834967939,
      "learning_rate": 5e-06,
      "loss": 0.4995,
      "step": 60
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6619830987803818,
      "learning_rate": 5e-06,
      "loss": 0.4931,
      "step": 70
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.5173934698104858,
      "eval_runtime": 3.7578,
      "eval_samples_per_second": 245.089,
      "eval_steps_per_second": 1.064,
      "step": 70
    },
    {
      "epoch": 2.2857142857142856,
      "grad_norm": 0.7326230444692339,
      "learning_rate": 5e-06,
      "loss": 0.461,
      "step": 80
    },
    {
      "epoch": 2.571428571428571,
      "grad_norm": 0.47640025795431173,
      "learning_rate": 5e-06,
      "loss": 0.4542,
      "step": 90
    },
    {
      "epoch": 2.857142857142857,
      "grad_norm": 0.5653219375676357,
      "learning_rate": 5e-06,
      "loss": 0.4527,
      "step": 100
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.5122517347335815,
      "eval_runtime": 3.6345,
      "eval_samples_per_second": 253.402,
      "eval_steps_per_second": 1.101,
      "step": 105
    },
    {
      "epoch": 3.0,
      "step": 105,
      "total_flos": 175878910771200.0,
      "train_loss": 0.5216304597400484,
      "train_runtime": 1018.6655,
      "train_samples_per_second": 51.529,
      "train_steps_per_second": 0.103
    }
  ],
  "logging_steps": 10,
  "max_steps": 105,
  "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": 175878910771200.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}