last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.75,
  "eval_steps": 500,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "learning_rate": 4.590984974958264e-05,
      "loss": 0.3734,
      "step": 500
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.089206762611866,
      "eval_runtime": 2506.8412,
      "eval_samples_per_second": 0.199,
      "eval_steps_per_second": 0.05,
      "step": 500
    },
    {
      "epoch": 0.5,
      "learning_rate": 4.1736227045075125e-05,
      "loss": 0.0979,
      "step": 1000
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.06451611965894699,
      "eval_runtime": 2530.4676,
      "eval_samples_per_second": 0.198,
      "eval_steps_per_second": 0.049,
      "step": 1000
    },
    {
      "epoch": 0.75,
      "learning_rate": 3.756260434056762e-05,
      "loss": 0.1096,
      "step": 1500
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.041419126093387604,
      "eval_runtime": 2505.8009,
      "eval_samples_per_second": 0.2,
      "eval_steps_per_second": 0.05,
      "step": 1500
    }
  ],
  "logging_steps": 500,
  "max_steps": 6000,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 6501313806336000.0,
  "trial_name": null,
  "trial_params": null
}