trainer_state.json

{
  "best_metric": 0.31285223364830017,
  "best_model_checkpoint": "hBERTv2_qqp/checkpoint-4266",
  "epoch": 8.0,
  "global_step": 11376,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 4.900070323488045e-05,
      "loss": 0.4179,
      "step": 1422
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8251545881770962,
      "eval_combined_score": 0.8083731607119745,
      "eval_f1": 0.7915917332468528,
      "eval_loss": 0.38297852873802185,
      "eval_runtime": 47.1453,
      "eval_samples_per_second": 857.561,
      "eval_steps_per_second": 3.351,
      "step": 1422
    },
    {
      "epoch": 2.0,
      "learning_rate": 4.800070323488045e-05,
      "loss": 0.2978,
      "step": 2844
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8357160524363096,
      "eval_combined_score": 0.8131408666041042,
      "eval_f1": 0.7905656807718988,
      "eval_loss": 0.35074833035469055,
      "eval_runtime": 47.1716,
      "eval_samples_per_second": 857.083,
      "eval_steps_per_second": 3.349,
      "step": 2844
    },
    {
      "epoch": 3.0,
      "learning_rate": 4.700070323488045e-05,
      "loss": 0.2318,
      "step": 4266
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8651001731387583,
      "eval_combined_score": 0.8405646585183773,
      "eval_f1": 0.8160291438979962,
      "eval_loss": 0.31285223364830017,
      "eval_runtime": 47.2025,
      "eval_samples_per_second": 856.522,
      "eval_steps_per_second": 3.347,
      "step": 4266
    },
    {
      "epoch": 4.0,
      "learning_rate": 4.600070323488046e-05,
      "loss": 0.1765,
      "step": 5688
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8700469948058372,
      "eval_combined_score": 0.8514036858419307,
      "eval_f1": 0.8327603768780241,
      "eval_loss": 0.35401350259780884,
      "eval_runtime": 47.169,
      "eval_samples_per_second": 857.131,
      "eval_steps_per_second": 3.35,
      "step": 5688
    },
    {
      "epoch": 5.0,
      "learning_rate": 4.500070323488045e-05,
      "loss": 0.1305,
      "step": 7110
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8733860994311156,
      "eval_combined_score": 0.8500331546757744,
      "eval_f1": 0.8266802099204335,
      "eval_loss": 0.4275738596916199,
      "eval_runtime": 47.244,
      "eval_samples_per_second": 855.769,
      "eval_steps_per_second": 3.344,
      "step": 7110
    },
    {
      "epoch": 6.0,
      "learning_rate": 4.40014064697609e-05,
      "loss": 0.1003,
      "step": 8532
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8748454118229038,
      "eval_combined_score": 0.8520305093406406,
      "eval_f1": 0.8292156068583774,
      "eval_loss": 0.40782323479652405,
      "eval_runtime": 47.2138,
      "eval_samples_per_second": 856.317,
      "eval_steps_per_second": 3.346,
      "step": 8532
    },
    {
      "epoch": 7.0,
      "learning_rate": 4.30014064697609e-05,
      "loss": 0.0788,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.876675735839723,
      "eval_combined_score": 0.855563447307603,
      "eval_f1": 0.8344511587754831,
      "eval_loss": 0.4068982005119324,
      "eval_runtime": 47.2202,
      "eval_samples_per_second": 856.201,
      "eval_steps_per_second": 3.346,
      "step": 9954
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.200210970464135e-05,
      "loss": 0.0625,
      "step": 11376
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8760079149146673,
      "eval_combined_score": 0.8541156077803029,
      "eval_f1": 0.8322233006459385,
      "eval_loss": 0.4722978472709656,
      "eval_runtime": 46.9884,
      "eval_samples_per_second": 860.425,
      "eval_steps_per_second": 3.363,
      "step": 11376
    },
    {
      "epoch": 8.0,
      "step": 11376,
      "total_flos": 3.625461609146614e+17,
      "train_loss": 0.18701694115640075,
      "train_runtime": 10469.4901,
      "train_samples_per_second": 1737.649,
      "train_steps_per_second": 6.791
    }
  ],
  "max_steps": 71100,
  "num_train_epochs": 50,
  "total_flos": 3.625461609146614e+17,
  "trial_name": null,
  "trial_params": null
}