last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.108695652173913,
  "eval_steps": 1,
  "global_step": 3,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.34782608695652173,
      "grad_norm": 48.1169319152832,
      "learning_rate": 1e-05,
      "loss": 19.6803,
      "step": 1
    },
    {
      "epoch": 0.34782608695652173,
      "eval_loss": 19.403636932373047,
      "eval_runtime": 173.3762,
      "eval_samples_per_second": 0.029,
      "eval_steps_per_second": 0.017,
      "step": 1
    },
    {
      "epoch": 0.6956521739130435,
      "grad_norm": 48.137481689453125,
      "learning_rate": 2e-05,
      "loss": 19.8041,
      "step": 2
    },
    {
      "epoch": 0.6956521739130435,
      "eval_loss": 19.008075714111328,
      "eval_runtime": 0.0898,
      "eval_samples_per_second": 55.692,
      "eval_steps_per_second": 33.415,
      "step": 2
    },
    {
      "epoch": 1.108695652173913,
      "grad_norm": 61.87843704223633,
      "learning_rate": 3e-05,
      "loss": 25.8279,
      "step": 3
    },
    {
      "epoch": 1.108695652173913,
      "eval_loss": 18.228900909423828,
      "eval_runtime": 0.0897,
      "eval_samples_per_second": 55.754,
      "eval_steps_per_second": 33.453,
      "step": 3
    }
  ],
  "logging_steps": 1,
  "max_steps": 3,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 612619230117888.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}