trainer_state.json

{
  "best_metric": 1.5944311618804932,
  "best_model_checkpoint": "/nfs/paraphrasing/output/bert-finetuned-para/checkpoint-1500",
  "epoch": 0.008791827117511562,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.9998827756384335e-05,
      "loss": 2.0282,
      "step": 500
    },
    {
      "epoch": 0.0,
      "eval_gen_len": 13.1456,
      "eval_loss": 1.6860297918319702,
      "eval_rouge1": 55.3995,
      "eval_rouge2": 32.0094,
      "eval_rougeL": 53.1553,
      "eval_rougeLsum": 53.1706,
      "eval_runtime": 50298.9522,
      "eval_samples_per_second": 27.136,
      "eval_steps_per_second": 0.424,
      "step": 500
    },
    {
      "epoch": 0.01,
      "learning_rate": 1.9997655512768664e-05,
      "loss": 1.8116,
      "step": 1000
    },
    {
      "epoch": 0.01,
      "eval_gen_len": 13.1187,
      "eval_loss": 1.6253910064697266,
      "eval_rouge1": 56.0224,
      "eval_rouge2": 32.6912,
      "eval_rougeL": 53.7965,
      "eval_rougeLsum": 53.8115,
      "eval_runtime": 111894.1395,
      "eval_samples_per_second": 12.198,
      "eval_steps_per_second": 0.191,
      "step": 1000
    },
    {
      "epoch": 0.01,
      "learning_rate": 1.9996483269152997e-05,
      "loss": 1.7615,
      "step": 1500
    },
    {
      "epoch": 0.01,
      "eval_gen_len": 13.0796,
      "eval_loss": 1.5944311618804932,
      "eval_rouge1": 56.3476,
      "eval_rouge2": 33.0715,
      "eval_rougeL": 54.1323,
      "eval_rougeLsum": 54.1475,
      "eval_runtime": 102094.2868,
      "eval_samples_per_second": 13.369,
      "eval_steps_per_second": 0.209,
      "step": 1500
    }
  ],
  "max_steps": 8530650,
  "num_train_epochs": 50,
  "total_flos": 4699507207962624.0,
  "trial_name": null,
  "trial_params": null
}