trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 8915,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.4032909870147705,
      "learning_rate": 4e-05,
      "loss": 3.5583,
      "step": 1783
    },
    {
      "epoch": 1.0,
      "eval_gen_len": 1.0,
      "eval_loss": 2.600463628768921,
      "eval_rouge1": 0.4848,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.4883,
      "eval_rougeLsum": 0.4866,
      "eval_runtime": 162.5937,
      "eval_samples_per_second": 4.613,
      "eval_steps_per_second": 0.148,
      "step": 1783
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.3332852125167847,
      "learning_rate": 3e-05,
      "loss": 3.0833,
      "step": 3566
    },
    {
      "epoch": 2.0,
      "eval_gen_len": 1.0,
      "eval_loss": 2.4166650772094727,
      "eval_rouge1": 0.4635,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.4663,
      "eval_rougeLsum": 0.4673,
      "eval_runtime": 160.1776,
      "eval_samples_per_second": 4.682,
      "eval_steps_per_second": 0.15,
      "step": 3566
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.7508715391159058,
      "learning_rate": 2e-05,
      "loss": 2.9538,
      "step": 5349
    },
    {
      "epoch": 3.0,
      "eval_gen_len": 1.0,
      "eval_loss": 2.343231439590454,
      "eval_rouge1": 0.5124,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.5075,
      "eval_rougeLsum": 0.5087,
      "eval_runtime": 163.6671,
      "eval_samples_per_second": 4.582,
      "eval_steps_per_second": 0.147,
      "step": 5349
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.3061054944992065,
      "learning_rate": 1e-05,
      "loss": 2.8804,
      "step": 7132
    },
    {
      "epoch": 4.0,
      "eval_gen_len": 1.0,
      "eval_loss": 2.2963712215423584,
      "eval_rouge1": 0.4384,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.4392,
      "eval_rougeLsum": 0.4405,
      "eval_runtime": 163.1032,
      "eval_samples_per_second": 4.598,
      "eval_steps_per_second": 0.147,
      "step": 7132
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.440184235572815,
      "learning_rate": 0.0,
      "loss": 2.8424,
      "step": 8915
    },
    {
      "epoch": 5.0,
      "eval_gen_len": 1.0,
      "eval_loss": 2.2596616744995117,
      "eval_rouge1": 0.4544,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.4545,
      "eval_rougeLsum": 0.4552,
      "eval_runtime": 162.5791,
      "eval_samples_per_second": 4.613,
      "eval_steps_per_second": 0.148,
      "step": 8915
    },
    {
      "epoch": 5.0,
      "step": 8915,
      "total_flos": 8.512744818393907e+16,
      "train_loss": 3.0636178162682275,
      "train_runtime": 5809.9425,
      "train_samples_per_second": 12.274,
      "train_steps_per_second": 1.534
    }
  ],
  "logging_steps": 500,
  "max_steps": 8915,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 8.512744818393907e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}