trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.008176347462061747,
  "eval_steps": 100,
  "global_step": 125,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 1.278611660003662,
      "learning_rate": 4.4e-07,
      "loss": 1.3731,
      "step": 25
    },
    {
      "epoch": 0.0,
      "grad_norm": 1.7951180934906006,
      "learning_rate": 9.400000000000001e-07,
      "loss": 1.8303,
      "step": 50
    },
    {
      "epoch": 0.0,
      "grad_norm": 0.9430514574050903,
      "learning_rate": 1.44e-06,
      "loss": 1.5785,
      "step": 75
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.380728006362915,
      "learning_rate": 1.9200000000000003e-06,
      "loss": 2.041,
      "step": 100
    },
    {
      "epoch": 0.01,
      "eval_loss": 1.6906120777130127,
      "eval_runtime": 2489.9511,
      "eval_samples_per_second": 1.535,
      "eval_steps_per_second": 1.535,
      "step": 100
    },
    {
      "epoch": 0.01,
      "grad_norm": 0.9432912468910217,
      "learning_rate": 2.42e-06,
      "loss": 1.4235,
      "step": 125
    }
  ],
  "logging_steps": 25,
  "max_steps": 15288,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 125,
  "total_flos": 1100373790703616.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}