trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.705985963344574,
  "best_model_checkpoint": "bert_base_rand_100_v1_wnli/checkpoint-6",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.281815528869629,
      "learning_rate": 4.9e-05,
      "loss": 0.6987,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.29577464788732394,
      "eval_loss": 0.7065361142158508,
      "eval_runtime": 0.0563,
      "eval_samples_per_second": 1260.489,
      "eval_steps_per_second": 17.753,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8516596555709839,
      "learning_rate": 4.8e-05,
      "loss": 0.6957,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5070422535211268,
      "eval_loss": 0.705985963344574,
      "eval_runtime": 0.0607,
      "eval_samples_per_second": 1168.94,
      "eval_steps_per_second": 16.464,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.8403956294059753,
      "learning_rate": 4.7e-05,
      "loss": 0.694,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.30985915492957744,
      "eval_loss": 0.724856972694397,
      "eval_runtime": 0.0664,
      "eval_samples_per_second": 1069.361,
      "eval_steps_per_second": 15.061,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.3989112377166748,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6932,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4084507042253521,
      "eval_loss": 0.7265625,
      "eval_runtime": 0.0564,
      "eval_samples_per_second": 1258.63,
      "eval_steps_per_second": 17.727,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.599137783050537,
      "learning_rate": 4.5e-05,
      "loss": 0.693,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.22535211267605634,
      "eval_loss": 0.747854471206665,
      "eval_runtime": 0.0548,
      "eval_samples_per_second": 1294.566,
      "eval_steps_per_second": 18.233,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.5501086711883545,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6904,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.2535211267605634,
      "eval_loss": 0.784496009349823,
      "eval_runtime": 0.0553,
      "eval_samples_per_second": 1283.64,
      "eval_steps_per_second": 18.079,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.6854292154312134,
      "learning_rate": 4.3e-05,
      "loss": 0.6908,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.19718309859154928,
      "eval_loss": 0.820972740650177,
      "eval_runtime": 0.0772,
      "eval_samples_per_second": 920.042,
      "eval_steps_per_second": 12.958,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 584764320537600.0,
      "train_loss": 0.6937073639460972,
      "train_runtime": 22.2272,
      "train_samples_per_second": 1428.433,
      "train_steps_per_second": 6.749
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "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": 584764320537600.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}