trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.8604651162790697,
  "eval_steps": 500,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.47,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.0437,
      "step": 500
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.5549370050430298,
      "eval_runtime": 2055.9752,
      "eval_samples_per_second": 0.681,
      "eval_steps_per_second": 0.085,
      "step": 500
    },
    {
      "epoch": 0.93,
      "learning_rate": 2.5e-05,
      "loss": 0.0006,
      "step": 1000
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.6001065969467163,
      "eval_runtime": 1468.9475,
      "eval_samples_per_second": 0.953,
      "eval_steps_per_second": 0.119,
      "step": 1000
    },
    {
      "epoch": 1.4,
      "learning_rate": 1.25e-05,
      "loss": 0.0005,
      "step": 1500
    },
    {
      "epoch": 1.4,
      "eval_loss": 0.6242989301681519,
      "eval_runtime": 199.8198,
      "eval_samples_per_second": 7.006,
      "eval_steps_per_second": 0.876,
      "step": 1500
    },
    {
      "epoch": 1.86,
      "learning_rate": 0.0,
      "loss": 0.0003,
      "step": 2000
    },
    {
      "epoch": 1.86,
      "eval_loss": 0.6335224509239197,
      "eval_runtime": 212.8986,
      "eval_samples_per_second": 6.576,
      "eval_steps_per_second": 0.822,
      "step": 2000
    },
    {
      "epoch": 1.86,
      "step": 2000,
      "total_flos": 4246941057024000.0,
      "train_loss": 0.01125018835812807,
      "train_runtime": 169289.6715,
      "train_samples_per_second": 0.095,
      "train_steps_per_second": 0.012
    }
  ],
  "logging_steps": 500,
  "max_steps": 2000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 4246941057024000.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}