checkpoint-600/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5305039787798409,
  "eval_steps": 100,
  "global_step": 600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.08841732979664015,
      "eval_accuracy": 0.1989687629116883,
      "eval_loss": 4.78662633895874,
      "eval_runtime": 26.1083,
      "eval_samples_per_second": 16.7,
      "eval_steps_per_second": 0.536,
      "step": 100
    },
    {
      "epoch": 0.1768346595932803,
      "eval_accuracy": 0.330860012697261,
      "eval_loss": 4.04963493347168,
      "eval_runtime": 26.2098,
      "eval_samples_per_second": 16.635,
      "eval_steps_per_second": 0.534,
      "step": 200
    },
    {
      "epoch": 0.26525198938992045,
      "eval_accuracy": 0.37791993303243177,
      "eval_loss": 3.652527332305908,
      "eval_runtime": 26.1862,
      "eval_samples_per_second": 16.65,
      "eval_steps_per_second": 0.535,
      "step": 300
    },
    {
      "epoch": 0.3536693191865606,
      "eval_accuracy": 0.42577323076228935,
      "eval_loss": 3.2409958839416504,
      "eval_runtime": 26.1526,
      "eval_samples_per_second": 16.671,
      "eval_steps_per_second": 0.535,
      "step": 400
    },
    {
      "epoch": 0.4420866489832007,
      "grad_norm": 195.0,
      "learning_rate": 4.263188918361332e-05,
      "loss": 3.9116,
      "step": 500
    },
    {
      "epoch": 0.4420866489832007,
      "eval_accuracy": 0.39117474036699323,
      "eval_loss": 3.6305172443389893,
      "eval_runtime": 26.2335,
      "eval_samples_per_second": 16.62,
      "eval_steps_per_second": 0.534,
      "step": 500
    },
    {
      "epoch": 0.5305039787798409,
      "eval_accuracy": 0.4406458382696409,
      "eval_loss": 3.177022695541382,
      "eval_runtime": 26.2979,
      "eval_samples_per_second": 16.579,
      "eval_steps_per_second": 0.532,
      "step": 600
    }
  ],
  "logging_steps": 500,
  "max_steps": 3393,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 200,
  "total_flos": 3.05500218458112e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}