trainer_state.json

{
  "best_metric": 0.691364049911499,
  "best_model_checkpoint": "bert_base_lda_5_v1_qnli/checkpoint-2870",
  "epoch": 12.0,
  "eval_steps": 500,
  "global_step": 4920,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.2550521492958069,
      "learning_rate": 0.00098,
      "loss": 0.7195,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6935492157936096,
      "eval_runtime": 3.531,
      "eval_samples_per_second": 1547.152,
      "eval_steps_per_second": 6.231,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.3675061762332916,
      "learning_rate": 0.00096,
      "loss": 0.6932,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6934226751327515,
      "eval_runtime": 3.5215,
      "eval_samples_per_second": 1551.347,
      "eval_steps_per_second": 6.247,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.12713126838207245,
      "learning_rate": 0.00094,
      "loss": 0.6932,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6932539343833923,
      "eval_runtime": 3.5229,
      "eval_samples_per_second": 1550.697,
      "eval_steps_per_second": 6.245,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.01812651753425598,
      "learning_rate": 0.00092,
      "loss": 0.6932,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 3.4831,
      "eval_samples_per_second": 1568.421,
      "eval_steps_per_second": 6.316,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.047646842896938324,
      "learning_rate": 0.0009000000000000001,
      "loss": 0.6932,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6934648156166077,
      "eval_runtime": 3.5307,
      "eval_samples_per_second": 1547.287,
      "eval_steps_per_second": 6.231,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.05455608665943146,
      "learning_rate": 0.00088,
      "loss": 0.6932,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 3.5763,
      "eval_samples_per_second": 1527.541,
      "eval_steps_per_second": 6.152,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.019881071522831917,
      "learning_rate": 0.00086,
      "loss": 0.6931,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.691364049911499,
      "eval_runtime": 3.6965,
      "eval_samples_per_second": 1477.89,
      "eval_steps_per_second": 5.952,
      "step": 2870
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.13051092624664307,
      "learning_rate": 0.00084,
      "loss": 0.693,
      "step": 3280
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6934226751327515,
      "eval_runtime": 3.4868,
      "eval_samples_per_second": 1566.758,
      "eval_steps_per_second": 6.309,
      "step": 3280
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.05455218255519867,
      "learning_rate": 0.00082,
      "loss": 0.6932,
      "step": 3690
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 3.5099,
      "eval_samples_per_second": 1556.447,
      "eval_steps_per_second": 6.268,
      "step": 3690
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.2007143646478653,
      "learning_rate": 0.0008,
      "loss": 0.6933,
      "step": 4100
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6936336159706116,
      "eval_runtime": 3.4934,
      "eval_samples_per_second": 1563.794,
      "eval_steps_per_second": 6.298,
      "step": 4100
    },
    {
      "epoch": 11.0,
      "grad_norm": 0.09045995026826859,
      "learning_rate": 0.0007800000000000001,
      "loss": 0.6935,
      "step": 4510
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 3.5482,
      "eval_samples_per_second": 1539.66,
      "eval_steps_per_second": 6.2,
      "step": 4510
    },
    {
      "epoch": 12.0,
      "grad_norm": 0.014760163612663746,
      "learning_rate": 0.00076,
      "loss": 0.6932,
      "step": 4920
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6934226751327515,
      "eval_runtime": 3.5278,
      "eval_samples_per_second": 1548.578,
      "eval_steps_per_second": 6.236,
      "step": 4920
    },
    {
      "epoch": 12.0,
      "step": 4920,
      "total_flos": 1.6535424762943488e+17,
      "train_loss": 0.6953946369450267,
      "train_runtime": 2094.0003,
      "train_samples_per_second": 2501.026,
      "train_steps_per_second": 9.79
    }
  ],
  "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": 1.6535424762943488e+17,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}