checkpoint-109/trainer_state.json

{
  "best_metric": 0.13872365653514862,
  "best_model_checkpoint": "outputs/checkpoint-109",
  "epoch": 1.9863325740318907,
  "eval_steps": 500,
  "global_step": 109,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.36446469248291574,
      "grad_norm": 2.546876907348633,
      "learning_rate": 1.8000000000000001e-06,
      "loss": 0.373,
      "step": 20
    },
    {
      "epoch": 0.7289293849658315,
      "grad_norm": 1.2323638200759888,
      "learning_rate": 3.8000000000000005e-06,
      "loss": 0.2883,
      "step": 40
    },
    {
      "epoch": 0.9840546697038725,
      "eval_loss": 0.1884891539812088,
      "eval_runtime": 48.9142,
      "eval_samples_per_second": 3.435,
      "eval_steps_per_second": 0.429,
      "step": 54
    },
    {
      "epoch": 1.0933940774487472,
      "grad_norm": 0.5503110289573669,
      "learning_rate": 5.8e-06,
      "loss": 0.1829,
      "step": 60
    },
    {
      "epoch": 1.4578587699316627,
      "grad_norm": 0.3597412109375,
      "learning_rate": 7.800000000000002e-06,
      "loss": 0.1368,
      "step": 80
    },
    {
      "epoch": 1.8223234624145785,
      "grad_norm": 1.4852937459945679,
      "learning_rate": 9.800000000000001e-06,
      "loss": 0.1263,
      "step": 100
    },
    {
      "epoch": 1.9863325740318907,
      "eval_loss": 0.13872365653514862,
      "eval_runtime": 48.9202,
      "eval_samples_per_second": 3.434,
      "eval_steps_per_second": 0.429,
      "step": 109
    }
  ],
  "logging_steps": 20,
  "max_steps": 810,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 500,
  "total_flos": 4.112547618177024e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}