trainer_state.json

{
  "best_metric": 71.4437,
  "best_model_checkpoint": "bin/indosum-base-3/checkpoint-4460",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 4460,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.0639431476593018,
      "learning_rate": 0.0008,
      "loss": 1.1937,
      "step": 892
    },
    {
      "epoch": 1.0,
      "eval_gen_len": 90.984,
      "eval_loss": 0.8369650840759277,
      "eval_rouge1": 64.7751,
      "eval_rouge2": 56.7545,
      "eval_rougeL": 61.4956,
      "eval_rougeLsum": 63.81,
      "eval_runtime": 190.626,
      "eval_samples_per_second": 3.934,
      "eval_steps_per_second": 0.126,
      "step": 892
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.4769066572189331,
      "learning_rate": 0.0006,
      "loss": 0.6828,
      "step": 1784
    },
    {
      "epoch": 2.0,
      "eval_gen_len": 101.11733333333333,
      "eval_loss": 0.6911363005638123,
      "eval_rouge1": 69.9628,
      "eval_rouge2": 62.6338,
      "eval_rougeL": 66.8253,
      "eval_rougeLsum": 69.0763,
      "eval_runtime": 191.8081,
      "eval_samples_per_second": 3.91,
      "eval_steps_per_second": 0.125,
      "step": 1784
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.5286809802055359,
      "learning_rate": 0.0004,
      "loss": 0.4847,
      "step": 2676
    },
    {
      "epoch": 3.0,
      "eval_gen_len": 95.61333333333333,
      "eval_loss": 0.6692180037498474,
      "eval_rouge1": 69.9807,
      "eval_rouge2": 62.5614,
      "eval_rougeL": 66.7619,
      "eval_rougeLsum": 69.0683,
      "eval_runtime": 188.7212,
      "eval_samples_per_second": 3.974,
      "eval_steps_per_second": 0.127,
      "step": 2676
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.3024337887763977,
      "learning_rate": 0.0002,
      "loss": 0.3348,
      "step": 3568
    },
    {
      "epoch": 4.0,
      "eval_gen_len": 100.15466666666667,
      "eval_loss": 0.7028738856315613,
      "eval_rouge1": 70.9247,
      "eval_rouge2": 63.6191,
      "eval_rougeL": 67.7749,
      "eval_rougeLsum": 70.0079,
      "eval_runtime": 191.1263,
      "eval_samples_per_second": 3.924,
      "eval_steps_per_second": 0.126,
      "step": 3568
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.5054429173469543,
      "learning_rate": 0.0,
      "loss": 0.1988,
      "step": 4460
    },
    {
      "epoch": 5.0,
      "eval_gen_len": 98.66666666666667,
      "eval_loss": 0.7689540982246399,
      "eval_rouge1": 71.4437,
      "eval_rouge2": 64.1873,
      "eval_rougeL": 68.2379,
      "eval_rougeLsum": 70.5264,
      "eval_runtime": 192.0358,
      "eval_samples_per_second": 3.906,
      "eval_steps_per_second": 0.125,
      "step": 4460
    },
    {
      "epoch": 5.0,
      "step": 4460,
      "total_flos": 4.886428467068928e+16,
      "train_loss": 0.5789496143837146,
      "train_runtime": 5011.6095,
      "train_samples_per_second": 14.239,
      "train_steps_per_second": 0.89
    }
  ],
  "logging_steps": 500,
  "max_steps": 4460,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 4.886428467068928e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}