trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 30.0,
  "eval_steps": 8,
  "global_step": 60,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5,
      "learning_rate": 0,
      "loss": 3.3777,
      "step": 1
    },
    {
      "epoch": 0.5,
      "eval_loss": 4.1841230392456055,
      "eval_runtime": 0.0543,
      "eval_samples_per_second": 18.427,
      "eval_steps_per_second": 18.427,
      "step": 1
    },
    {
      "epoch": 4.0,
      "learning_rate": 4.935897435897436e-05,
      "loss": 3.0227,
      "step": 8
    },
    {
      "epoch": 4.0,
      "eval_loss": 1.9906407594680786,
      "eval_runtime": 0.0434,
      "eval_samples_per_second": 23.036,
      "eval_steps_per_second": 23.036,
      "step": 8
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.423076923076923e-05,
      "loss": 1.4343,
      "step": 16
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.4215693175792694,
      "eval_runtime": 0.0427,
      "eval_samples_per_second": 23.407,
      "eval_steps_per_second": 23.407,
      "step": 16
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.7243,
      "step": 24
    },
    {
      "epoch": 12.0,
      "eval_loss": 0.45053091645240784,
      "eval_runtime": 0.0425,
      "eval_samples_per_second": 23.522,
      "eval_steps_per_second": 23.522,
      "step": 24
    },
    {
      "epoch": 16.0,
      "learning_rate": 3.653846153846154e-05,
      "loss": 0.3808,
      "step": 32
    },
    {
      "epoch": 16.0,
      "eval_loss": 0.1258782297372818,
      "eval_runtime": 0.0427,
      "eval_samples_per_second": 23.394,
      "eval_steps_per_second": 23.394,
      "step": 32
    },
    {
      "epoch": 20.0,
      "learning_rate": 3.141025641025641e-05,
      "loss": 0.1583,
      "step": 40
    },
    {
      "epoch": 20.0,
      "eval_loss": 0.09587076306343079,
      "eval_runtime": 0.0435,
      "eval_samples_per_second": 23.007,
      "eval_steps_per_second": 23.007,
      "step": 40
    },
    {
      "epoch": 24.0,
      "learning_rate": 2.6282051282051285e-05,
      "loss": 0.0985,
      "step": 48
    },
    {
      "epoch": 24.0,
      "eval_loss": 0.08991015702486038,
      "eval_runtime": 0.0433,
      "eval_samples_per_second": 23.102,
      "eval_steps_per_second": 23.102,
      "step": 48
    },
    {
      "epoch": 28.0,
      "learning_rate": 2.1153846153846154e-05,
      "loss": 0.0549,
      "step": 56
    },
    {
      "epoch": 28.0,
      "eval_loss": 0.04217328876256943,
      "eval_runtime": 0.0433,
      "eval_samples_per_second": 23.103,
      "eval_steps_per_second": 23.103,
      "step": 56
    }
  ],
  "logging_steps": 8,
  "max_steps": 80,
  "num_train_epochs": 40,
  "save_steps": 20,
  "total_flos": 4871465287548928.0,
  "trial_name": null,
  "trial_params": null
}