trainer_state.json

{
  "best_metric": 0.48353341221809387,
  "best_model_checkpoint": "bert_base_lda_5_v1_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.9573304653167725,
      "learning_rate": 4.9e-05,
      "loss": 0.6432,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7056562328390994,
      "eval_loss": 0.5685960650444031,
      "eval_runtime": 3.5298,
      "eval_samples_per_second": 1547.692,
      "eval_steps_per_second": 6.233,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.638329267501831,
      "learning_rate": 4.8e-05,
      "loss": 0.514,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7721032399780341,
      "eval_loss": 0.48353341221809387,
      "eval_runtime": 3.544,
      "eval_samples_per_second": 1541.473,
      "eval_steps_per_second": 6.208,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.854526042938232,
      "learning_rate": 4.7e-05,
      "loss": 0.3845,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7651473549331869,
      "eval_loss": 0.5230943560600281,
      "eval_runtime": 3.5214,
      "eval_samples_per_second": 1551.356,
      "eval_steps_per_second": 6.247,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 8.325188636779785,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.2671,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7607541643785466,
      "eval_loss": 0.6301699280738831,
      "eval_runtime": 3.5684,
      "eval_samples_per_second": 1530.923,
      "eval_steps_per_second": 6.165,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.31276273727417,
      "learning_rate": 4.5e-05,
      "loss": 0.1749,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7634999084751968,
      "eval_loss": 0.7120682597160339,
      "eval_runtime": 3.6011,
      "eval_samples_per_second": 1517.049,
      "eval_steps_per_second": 6.109,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 8.96873950958252,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1231,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7618524620172067,
      "eval_loss": 0.7853891849517822,
      "eval_runtime": 3.5353,
      "eval_samples_per_second": 1545.281,
      "eval_steps_per_second": 6.223,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.691504001617432,
      "learning_rate": 4.3e-05,
      "loss": 0.0932,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7428153029470986,
      "eval_loss": 0.9599586129188538,
      "eval_runtime": 3.5825,
      "eval_samples_per_second": 1524.904,
      "eval_steps_per_second": 6.141,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.3142849300796562,
      "train_runtime": 1232.997,
      "train_samples_per_second": 4247.496,
      "train_steps_per_second": 16.626
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 5
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 9.645664445050368e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}