trainer_state.json

{
  "best_global_step": 2844,
  "best_metric": 0.3768726885318756,
  "best_model_checkpoint": "bert_base_km_5_v1_qqp/checkpoint-2844",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 9954,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.0276873111724854,
      "learning_rate": 4.9e-05,
      "loss": 0.4663,
      "step": 1422
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8001731387583477,
      "eval_combined_score": 0.7534249856736952,
      "eval_f1": 0.7066768325890426,
      "eval_loss": 0.4186016321182251,
      "eval_runtime": 25.8374,
      "eval_samples_per_second": 1564.788,
      "eval_steps_per_second": 6.115,
      "step": 1422
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.805953025817871,
      "learning_rate": 4.8e-05,
      "loss": 0.3533,
      "step": 2844
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8282958199356913,
      "eval_combined_score": 0.7974207106135276,
      "eval_f1": 0.766545601291364,
      "eval_loss": 0.3768726885318756,
      "eval_runtime": 25.5089,
      "eval_samples_per_second": 1584.937,
      "eval_steps_per_second": 6.194,
      "step": 2844
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.3630292415618896,
      "learning_rate": 4.7e-05,
      "loss": 0.2678,
      "step": 4266
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.829656195894138,
      "eval_combined_score": 0.8043851716184504,
      "eval_f1": 0.7791141473427627,
      "eval_loss": 0.39894676208496094,
      "eval_runtime": 25.5009,
      "eval_samples_per_second": 1585.435,
      "eval_steps_per_second": 6.196,
      "step": 4266
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.226913928985596,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.2011,
      "step": 5688
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8425179322285432,
      "eval_combined_score": 0.8098659852541995,
      "eval_f1": 0.7772140382798558,
      "eval_loss": 0.4219973683357239,
      "eval_runtime": 25.554,
      "eval_samples_per_second": 1582.141,
      "eval_steps_per_second": 6.183,
      "step": 5688
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.4161477088928223,
      "learning_rate": 4.5e-05,
      "loss": 0.1514,
      "step": 7110
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8437299035369775,
      "eval_combined_score": 0.8124206613028568,
      "eval_f1": 0.7811114190687362,
      "eval_loss": 0.49849799275398254,
      "eval_runtime": 25.5419,
      "eval_samples_per_second": 1582.892,
      "eval_steps_per_second": 6.186,
      "step": 7110
    },
    {
      "epoch": 6.0,
      "grad_norm": 5.633938789367676,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1188,
      "step": 8532
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8442493198120208,
      "eval_combined_score": 0.812934439687526,
      "eval_f1": 0.7816195595630311,
      "eval_loss": 0.5644199252128601,
      "eval_runtime": 25.5877,
      "eval_samples_per_second": 1580.054,
      "eval_steps_per_second": 6.175,
      "step": 8532
    },
    {
      "epoch": 7.0,
      "grad_norm": 5.966865539550781,
      "learning_rate": 4.3e-05,
      "loss": 0.0964,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8413306950284443,
      "eval_combined_score": 0.8130563965966524,
      "eval_f1": 0.7847820981648606,
      "eval_loss": 0.6045879125595093,
      "eval_runtime": 25.4437,
      "eval_samples_per_second": 1588.997,
      "eval_steps_per_second": 6.21,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "step": 9954,
      "total_flos": 3.3506166766980096e+17,
      "train_loss": 0.23642058063037236,
      "train_runtime": 4310.8303,
      "train_samples_per_second": 4220.138,
      "train_steps_per_second": 16.493
    }
  ],
  "logging_steps": 1,
  "max_steps": 71100,
  "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": 3.3506166766980096e+17,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}