checkpoint-16/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.0005333333333333334,
  "eval_steps": 2000,
  "global_step": 16,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 3.3333333333333335e-05,
      "eval_accuracy": 0.016437927663734114,
      "eval_loss": 10.702861785888672,
      "eval_runtime": 53.5239,
      "eval_samples_per_second": 93.416,
      "eval_steps_per_second": 2.933,
      "step": 1
    },
    {
      "epoch": 6.666666666666667e-05,
      "eval_accuracy": 0.049564027370478984,
      "eval_loss": 10.533119201660156,
      "eval_runtime": 53.3666,
      "eval_samples_per_second": 93.692,
      "eval_steps_per_second": 2.942,
      "step": 2
    },
    {
      "epoch": 0.00013333333333333334,
      "eval_accuracy": 0.05325317693059629,
      "eval_loss": 10.302159309387207,
      "eval_runtime": 53.3789,
      "eval_samples_per_second": 93.67,
      "eval_steps_per_second": 2.941,
      "step": 4
    },
    {
      "epoch": 0.0002666666666666667,
      "eval_accuracy": 0.053565591397849464,
      "eval_loss": 10.023477554321289,
      "eval_runtime": 53.6667,
      "eval_samples_per_second": 93.168,
      "eval_steps_per_second": 2.925,
      "step": 8
    },
    {
      "epoch": 0.0005333333333333334,
      "eval_accuracy": 0.06345063538611925,
      "eval_loss": 9.653569221496582,
      "eval_runtime": 53.2342,
      "eval_samples_per_second": 93.925,
      "eval_steps_per_second": 2.949,
      "step": 16
    }
  ],
  "logging_steps": 2000,
  "max_steps": 30000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 140522739990528.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}