trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 50.0,
  "eval_steps": 300,
  "global_step": 1700,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 8.82,
      "eval_bleu": 0.2892,
      "eval_gen_len": 127.2388,
      "eval_loss": 5.496266841888428,
      "eval_runtime": 169.1996,
      "eval_samples_per_second": 0.396,
      "eval_steps_per_second": 0.03,
      "step": 300
    },
    {
      "epoch": 14.71,
      "learning_rate": 4.328571428571429e-05,
      "loss": 8.4505,
      "step": 500
    },
    {
      "epoch": 17.65,
      "eval_bleu": 18.0955,
      "eval_gen_len": 12.7761,
      "eval_loss": 5.678420066833496,
      "eval_runtime": 4.5299,
      "eval_samples_per_second": 14.791,
      "eval_steps_per_second": 1.104,
      "step": 600
    },
    {
      "epoch": 26.47,
      "eval_bleu": 19.7871,
      "eval_gen_len": 13.6119,
      "eval_loss": 5.940618991851807,
      "eval_runtime": 4.4489,
      "eval_samples_per_second": 15.06,
      "eval_steps_per_second": 1.124,
      "step": 900
    },
    {
      "epoch": 29.41,
      "learning_rate": 2.542857142857143e-05,
      "loss": 0.4947,
      "step": 1000
    },
    {
      "epoch": 35.29,
      "eval_bleu": 16.6303,
      "eval_gen_len": 13.209,
      "eval_loss": 6.220270156860352,
      "eval_runtime": 4.5215,
      "eval_samples_per_second": 14.818,
      "eval_steps_per_second": 1.106,
      "step": 1200
    },
    {
      "epoch": 44.12,
      "learning_rate": 7.571428571428572e-06,
      "loss": 0.0447,
      "step": 1500
    },
    {
      "epoch": 44.12,
      "eval_bleu": 19.9934,
      "eval_gen_len": 14.0448,
      "eval_loss": 6.294175148010254,
      "eval_runtime": 4.5277,
      "eval_samples_per_second": 14.798,
      "eval_steps_per_second": 1.104,
      "step": 1500
    },
    {
      "epoch": 50.0,
      "step": 1700,
      "total_flos": 5.89459296878592e+16,
      "train_loss": 2.645088053380742,
      "train_runtime": 1356.8639,
      "train_samples_per_second": 19.862,
      "train_steps_per_second": 1.253
    }
  ],
  "logging_steps": 500,
  "max_steps": 1700,
  "num_train_epochs": 50,
  "save_steps": 300,
  "total_flos": 5.89459296878592e+16,
  "trial_name": null,
  "trial_params": null
}