checkpoint-60/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.8979591836734695,
  "eval_steps": 20,
  "global_step": 60,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.82,
      "grad_norm": 0.1748005896806717,
      "learning_rate": 2.1e-05,
      "loss": 4.8309,
      "step": 10
    },
    {
      "epoch": 1.63,
      "grad_norm": 0.19310609996318817,
      "learning_rate": 3.2999999999999996e-05,
      "loss": 5.0159,
      "step": 20
    },
    {
      "epoch": 1.63,
      "eval_loss": NaN,
      "eval_runtime": 53.474,
      "eval_samples_per_second": 7.237,
      "eval_steps_per_second": 0.916,
      "step": 20
    },
    {
      "epoch": 2.45,
      "grad_norm": NaN,
      "learning_rate": 4.2e-05,
      "loss": 4.8025,
      "step": 30
    },
    {
      "epoch": 3.27,
      "grad_norm": NaN,
      "learning_rate": 5.399999999999999e-05,
      "loss": 4.1014,
      "step": 40
    },
    {
      "epoch": 3.27,
      "eval_loss": NaN,
      "eval_runtime": 53.4777,
      "eval_samples_per_second": 7.237,
      "eval_steps_per_second": 0.916,
      "step": 40
    },
    {
      "epoch": 4.08,
      "grad_norm": 0.0,
      "learning_rate": 7.199999999999999e-05,
      "loss": 4.7645,
      "step": 50
    },
    {
      "epoch": 4.9,
      "grad_norm": 0.0,
      "learning_rate": 9.3e-05,
      "loss": 4.3747,
      "step": 60
    },
    {
      "epoch": 4.9,
      "eval_loss": NaN,
      "eval_runtime": 53.6778,
      "eval_samples_per_second": 7.21,
      "eval_steps_per_second": 0.913,
      "step": 60
    }
  ],
  "logging_steps": 10,
  "max_steps": 400,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 34,
  "save_steps": 20,
  "total_flos": 1.5222755513401344e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}