last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.25,
  "eval_steps": 24,
  "global_step": 24,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.010416666666666666,
      "eval_loss": 1.9108808040618896,
      "eval_runtime": 3.3051,
      "eval_samples_per_second": 49.016,
      "eval_steps_per_second": 6.354,
      "step": 1
    },
    {
      "epoch": 0.03125,
      "grad_norm": 0.41242459416389465,
      "learning_rate": 1.5e-05,
      "loss": 1.8948,
      "step": 3
    },
    {
      "epoch": 0.0625,
      "grad_norm": 0.5248602032661438,
      "learning_rate": 3e-05,
      "loss": 1.9034,
      "step": 6
    },
    {
      "epoch": 0.09375,
      "grad_norm": 0.4381963312625885,
      "learning_rate": 4.5e-05,
      "loss": 1.8402,
      "step": 9
    },
    {
      "epoch": 0.125,
      "grad_norm": 0.3652794063091278,
      "learning_rate": 4.99936149886953e-05,
      "loss": 1.7817,
      "step": 12
    },
    {
      "epoch": 0.15625,
      "grad_norm": 0.30795106291770935,
      "learning_rate": 4.996010259708474e-05,
      "loss": 1.7055,
      "step": 15
    },
    {
      "epoch": 0.1875,
      "grad_norm": 0.35701417922973633,
      "learning_rate": 4.989790503518888e-05,
      "loss": 1.7031,
      "step": 18
    },
    {
      "epoch": 0.21875,
      "grad_norm": 0.3083578944206238,
      "learning_rate": 4.980709378298851e-05,
      "loss": 1.6675,
      "step": 21
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.2675102651119232,
      "learning_rate": 4.9687773204487064e-05,
      "loss": 1.6823,
      "step": 24
    },
    {
      "epoch": 0.25,
      "eval_loss": 1.6583374738693237,
      "eval_runtime": 3.3211,
      "eval_samples_per_second": 48.78,
      "eval_steps_per_second": 6.323,
      "step": 24
    }
  ],
  "logging_steps": 3,
  "max_steps": 288,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 24,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 5439527023804416.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}