trainer_state.json

{
  "best_metric": 19.886,
  "best_model_checkpoint": "bin/liputan6-base/checkpoint-189",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 189,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 5.871487617492676,
      "learning_rate": 0.0006666666666666666,
      "loss": 3.7824,
      "step": 63
    },
    {
      "epoch": 1.0,
      "eval_gen_len": 31.998,
      "eval_loss": 3.822542190551758,
      "eval_rouge1": 16.8681,
      "eval_rouge2": 5.1649,
      "eval_rougeL": 14.9393,
      "eval_rougeLsum": 15.6881,
      "eval_runtime": 235.6514,
      "eval_samples_per_second": 4.244,
      "eval_steps_per_second": 0.136,
      "step": 63
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.992724895477295,
      "learning_rate": 0.0003333333333333333,
      "loss": 2.0598,
      "step": 126
    },
    {
      "epoch": 2.0,
      "eval_gen_len": 38.822,
      "eval_loss": 3.8557004928588867,
      "eval_rouge1": 18.6028,
      "eval_rouge2": 5.9301,
      "eval_rougeL": 15.9454,
      "eval_rougeLsum": 17.1143,
      "eval_runtime": 247.4796,
      "eval_samples_per_second": 4.041,
      "eval_steps_per_second": 0.129,
      "step": 126
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.042899131774902,
      "learning_rate": 0.0,
      "loss": 0.945,
      "step": 189
    },
    {
      "epoch": 3.0,
      "eval_gen_len": 30.875,
      "eval_loss": 4.205010414123535,
      "eval_rouge1": 19.886,
      "eval_rouge2": 6.0542,
      "eval_rougeL": 16.8703,
      "eval_rougeLsum": 18.2573,
      "eval_runtime": 188.2411,
      "eval_samples_per_second": 5.312,
      "eval_steps_per_second": 0.17,
      "step": 189
    },
    {
      "epoch": 3.0,
      "step": 189,
      "total_flos": 2054272057344000.0,
      "train_loss": 2.2623963128952753,
      "train_runtime": 857.0722,
      "train_samples_per_second": 3.5,
      "train_steps_per_second": 0.221
    }
  ],
  "logging_steps": 500,
  "max_steps": 189,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 2054272057344000.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}