trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5,
      "grad_norm": 18.081775665283203,
      "learning_rate": 1e-05,
      "loss": 1.1002,
      "step": 1
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.2711809873580933,
      "eval_runtime": 5.1274,
      "eval_samples_per_second": 9.752,
      "eval_steps_per_second": 0.39,
      "step": 2
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.8800549507141113,
      "eval_runtime": 5.1265,
      "eval_samples_per_second": 9.753,
      "eval_steps_per_second": 0.39,
      "step": 4
    },
    {
      "epoch": 2.5,
      "grad_norm": 19.409046173095703,
      "learning_rate": 8.94570254698197e-06,
      "loss": 1.0859,
      "step": 5
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.6101047992706299,
      "eval_runtime": 5.1466,
      "eval_samples_per_second": 9.715,
      "eval_steps_per_second": 0.389,
      "step": 6
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.44190600514411926,
      "eval_runtime": 5.1566,
      "eval_samples_per_second": 9.696,
      "eval_steps_per_second": 0.388,
      "step": 8
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.709940433502197,
      "learning_rate": 5.412896727361663e-06,
      "loss": 0.5465,
      "step": 10
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.26203182339668274,
      "eval_runtime": 5.1422,
      "eval_samples_per_second": 9.724,
      "eval_steps_per_second": 0.389,
      "step": 10
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.14132770895957947,
      "eval_runtime": 5.1462,
      "eval_samples_per_second": 9.716,
      "eval_steps_per_second": 0.389,
      "step": 12
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.08078073710203171,
      "eval_runtime": 5.1421,
      "eval_samples_per_second": 9.724,
      "eval_steps_per_second": 0.389,
      "step": 14
    },
    {
      "epoch": 7.5,
      "grad_norm": 1.5183640718460083,
      "learning_rate": 1.6135921418712959e-06,
      "loss": 0.1621,
      "step": 15
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.05351748317480087,
      "eval_runtime": 5.1503,
      "eval_samples_per_second": 9.708,
      "eval_steps_per_second": 0.388,
      "step": 16
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.042922407388687134,
      "eval_runtime": 5.1564,
      "eval_samples_per_second": 9.697,
      "eval_steps_per_second": 0.388,
      "step": 18
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.851169228553772,
      "learning_rate": 0.0,
      "loss": 0.0504,
      "step": 20
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.04061064124107361,
      "eval_runtime": 5.054,
      "eval_samples_per_second": 9.893,
      "eval_steps_per_second": 0.396,
      "step": 20
    },
    {
      "epoch": 10.0,
      "step": 20,
      "total_flos": 1.1184101603357491e+17,
      "train_loss": 0.46195218414068223,
      "train_runtime": 339.5674,
      "train_samples_per_second": 1.472,
      "train_steps_per_second": 0.059
    }
  ],
  "logging_steps": 5,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1184101603357491e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}