checkpoint-100/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.34965034965034963,
  "eval_steps": 25,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09,
      "grad_norm": 6.918973445892334,
      "learning_rate": 1.893939393939394e-05,
      "loss": 3.9448,
      "step": 25
    },
    {
      "epoch": 0.09,
      "eval_loss": 3.8408708572387695,
      "eval_runtime": 6.6099,
      "eval_samples_per_second": 21.786,
      "eval_steps_per_second": 2.723,
      "step": 25
    },
    {
      "epoch": 0.17,
      "grad_norm": 5.2677106857299805,
      "learning_rate": 1.2626262626262628e-05,
      "loss": 3.9142,
      "step": 50
    },
    {
      "epoch": 0.17,
      "eval_loss": 3.7116785049438477,
      "eval_runtime": 6.775,
      "eval_samples_per_second": 21.255,
      "eval_steps_per_second": 2.657,
      "step": 50
    },
    {
      "epoch": 0.26,
      "grad_norm": 6.679225921630859,
      "learning_rate": 6.313131313131314e-06,
      "loss": 3.6859,
      "step": 75
    },
    {
      "epoch": 0.26,
      "eval_loss": 3.639355182647705,
      "eval_runtime": 7.1093,
      "eval_samples_per_second": 20.255,
      "eval_steps_per_second": 2.532,
      "step": 75
    },
    {
      "epoch": 0.35,
      "grad_norm": 4.882274627685547,
      "learning_rate": 0.0,
      "loss": 3.7328,
      "step": 100
    },
    {
      "epoch": 0.35,
      "eval_loss": 3.6143908500671387,
      "eval_runtime": 7.1262,
      "eval_samples_per_second": 20.207,
      "eval_steps_per_second": 2.526,
      "step": 100
    }
  ],
  "logging_steps": 25,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 25,
  "total_flos": 54349057228800.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}