checkpoint-108/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 25,
  "global_step": 108,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.6944444444444444,
      "grad_norm": 1.3425393104553223,
      "learning_rate": 0.0005,
      "loss": 1.553,
      "step": 25
    },
    {
      "epoch": 0.6944444444444444,
      "eval_loss": 0.8388024568557739,
      "eval_runtime": 107.6689,
      "eval_samples_per_second": 2.015,
      "eval_steps_per_second": 0.26,
      "step": 25
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 0.6725547909736633,
      "learning_rate": 0.001,
      "loss": 0.508,
      "step": 50
    },
    {
      "epoch": 1.3888888888888888,
      "eval_loss": 0.44558215141296387,
      "eval_runtime": 105.6439,
      "eval_samples_per_second": 2.054,
      "eval_steps_per_second": 0.265,
      "step": 50
    },
    {
      "epoch": 2.0833333333333335,
      "grad_norm": 0.5922499299049377,
      "learning_rate": 0.0005862068965517241,
      "loss": 0.3276,
      "step": 75
    },
    {
      "epoch": 2.0833333333333335,
      "eval_loss": 0.3984259068965912,
      "eval_runtime": 106.5674,
      "eval_samples_per_second": 2.036,
      "eval_steps_per_second": 0.263,
      "step": 75
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 0.8124176263809204,
      "learning_rate": 0.00015517241379310346,
      "loss": 0.16,
      "step": 100
    },
    {
      "epoch": 2.7777777777777777,
      "eval_loss": 0.39040103554725647,
      "eval_runtime": 107.0245,
      "eval_samples_per_second": 2.028,
      "eval_steps_per_second": 0.262,
      "step": 100
    }
  ],
  "logging_steps": 25,
  "max_steps": 108,
  "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": 1.8218014027776e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}