trainer_state.json

{
  "best_metric": 1.7082571983337402,
  "best_model_checkpoint": "output/ramil/checkpoint-33",
  "epoch": 1.0,
  "global_step": 33,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15,
      "learning_rate": 0.00012957411177772776,
      "loss": 2.2142,
      "step": 5
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0001083919039965842,
      "loss": 2.1038,
      "step": 10
    },
    {
      "epoch": 0.45,
      "learning_rate": 7.836279790554737e-05,
      "loss": 1.8934,
      "step": 15
    },
    {
      "epoch": 0.61,
      "learning_rate": 4.616313771642488e-05,
      "loss": 1.9451,
      "step": 20
    },
    {
      "epoch": 0.76,
      "learning_rate": 1.8951844985992186e-05,
      "loss": 1.8488,
      "step": 25
    },
    {
      "epoch": 0.91,
      "learning_rate": 2.778782010045488e-06,
      "loss": 1.8871,
      "step": 30
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.7082571983337402,
      "eval_runtime": 2.2102,
      "eval_samples_per_second": 22.623,
      "eval_steps_per_second": 3.167,
      "step": 33
    }
  ],
  "max_steps": 33,
  "num_train_epochs": 1,
  "total_flos": 34098610176000.0,
  "trial_name": null,
  "trial_params": null
}