trainer_state.json

{
  "best_global_step": 264,
  "best_metric": 0.4423932433128357,
  "best_model_checkpoint": "bert_base_rand_100_v2_sst2/checkpoint-264",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 1584,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 9.260363578796387,
      "learning_rate": 4.9e-05,
      "loss": 0.3852,
      "step": 264
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8061926605504587,
      "eval_loss": 0.4423932433128357,
      "eval_runtime": 0.5775,
      "eval_samples_per_second": 1509.959,
      "eval_steps_per_second": 6.926,
      "step": 264
    },
    {
      "epoch": 2.0,
      "grad_norm": 9.812104225158691,
      "learning_rate": 4.8e-05,
      "loss": 0.2193,
      "step": 528
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7970183486238532,
      "eval_loss": 0.5311356782913208,
      "eval_runtime": 0.57,
      "eval_samples_per_second": 1529.879,
      "eval_steps_per_second": 7.018,
      "step": 528
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.92247200012207,
      "learning_rate": 4.7e-05,
      "loss": 0.1569,
      "step": 792
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8084862385321101,
      "eval_loss": 0.5927096605300903,
      "eval_runtime": 0.5721,
      "eval_samples_per_second": 1524.336,
      "eval_steps_per_second": 6.992,
      "step": 792
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.917325019836426,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.1203,
      "step": 1056
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7981651376146789,
      "eval_loss": 0.5860992670059204,
      "eval_runtime": 0.6168,
      "eval_samples_per_second": 1413.795,
      "eval_steps_per_second": 6.485,
      "step": 1056
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.848937034606934,
      "learning_rate": 4.5e-05,
      "loss": 0.0926,
      "step": 1320
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8038990825688074,
      "eval_loss": 0.6796774864196777,
      "eval_runtime": 0.5696,
      "eval_samples_per_second": 1530.767,
      "eval_steps_per_second": 7.022,
      "step": 1320
    },
    {
      "epoch": 6.0,
      "grad_norm": 10.027762413024902,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.0721,
      "step": 1584
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7947247706422018,
      "eval_loss": 0.8250018358230591,
      "eval_runtime": 0.5705,
      "eval_samples_per_second": 1528.455,
      "eval_steps_per_second": 7.011,
      "step": 1584
    },
    {
      "epoch": 6.0,
      "step": 1584,
      "total_flos": 5.316079940232192e+16,
      "train_loss": 0.17438846886760057,
      "train_runtime": 669.727,
      "train_samples_per_second": 5028.094,
      "train_steps_per_second": 19.71
    }
  ],
  "logging_steps": 1,
  "max_steps": 13200,
  "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": 5.316079940232192e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}