checkpoint-2000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.7479248580166011,
  "eval_steps": 500,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.46,
      "grad_norm": 0.36198097467422485,
      "learning_rate": 6.0716454159077105e-05,
      "loss": 0.4455,
      "step": 500
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.6183120608329773,
      "eval_runtime": 195.3588,
      "eval_samples_per_second": 2.37,
      "eval_steps_per_second": 2.37,
      "step": 500
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.1898297369480133,
      "learning_rate": 0.00012143290831815421,
      "loss": 0.0995,
      "step": 1000
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.6644699573516846,
      "eval_runtime": 195.5381,
      "eval_samples_per_second": 2.368,
      "eval_steps_per_second": 2.368,
      "step": 1000
    },
    {
      "epoch": 1.37,
      "grad_norm": 0.25010770559310913,
      "learning_rate": 0.00018214936247723133,
      "loss": 0.0858,
      "step": 1500
    },
    {
      "epoch": 1.37,
      "eval_loss": 0.6735700368881226,
      "eval_runtime": 195.292,
      "eval_samples_per_second": 2.371,
      "eval_steps_per_second": 2.371,
      "step": 1500
    },
    {
      "epoch": 1.7479248580166011,
      "grad_norm": 0.12149681895971298,
      "learning_rate": 0.00019897628145050826,
      "loss": 0.0836,
      "step": 2000
    },
    {
      "epoch": 1.7479248580166011,
      "eval_loss": 0.2921118438243866,
      "eval_runtime": 49.9169,
      "eval_samples_per_second": 1.863,
      "eval_steps_per_second": 1.863,
      "step": 2000
    }
  ],
  "logging_steps": 500,
  "max_steps": 57200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "total_flos": 1.6859431125624422e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}