trainer_state.json

{
  "best_global_step": 2844,
  "best_metric": 0.3907022476196289,
  "best_model_checkpoint": "bert_base_rand_5_v2_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": 1.6056857109069824,
      "learning_rate": 4.9e-05,
      "loss": 0.4731,
      "step": 1422
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7941874845411823,
      "eval_combined_score": 0.7453337568611438,
      "eval_f1": 0.6964800291811052,
      "eval_loss": 0.42616578936576843,
      "eval_runtime": 25.954,
      "eval_samples_per_second": 1557.753,
      "eval_steps_per_second": 6.088,
      "step": 1422
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.522731304168701,
      "learning_rate": 4.8e-05,
      "loss": 0.3673,
      "step": 2844
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8173880781597823,
      "eval_combined_score": 0.7884536560304791,
      "eval_f1": 0.7595192339011758,
      "eval_loss": 0.3907022476196289,
      "eval_runtime": 25.6425,
      "eval_samples_per_second": 1576.677,
      "eval_steps_per_second": 6.162,
      "step": 2844
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.13094425201416,
      "learning_rate": 4.7e-05,
      "loss": 0.2918,
      "step": 4266
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.827900074202325,
      "eval_combined_score": 0.801258925414197,
      "eval_f1": 0.7746177766260689,
      "eval_loss": 0.40507155656814575,
      "eval_runtime": 25.329,
      "eval_samples_per_second": 1596.193,
      "eval_steps_per_second": 6.238,
      "step": 4266
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.6782093048095703,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.2294,
      "step": 5688
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8362849369280237,
      "eval_combined_score": 0.803416500425741,
      "eval_f1": 0.7705480639234582,
      "eval_loss": 0.4142085015773773,
      "eval_runtime": 25.5461,
      "eval_samples_per_second": 1582.629,
      "eval_steps_per_second": 6.185,
      "step": 5688
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.8276795148849487,
      "learning_rate": 4.5e-05,
      "loss": 0.1819,
      "step": 7110
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8399703190699975,
      "eval_combined_score": 0.8097486707460667,
      "eval_f1": 0.7795270224221359,
      "eval_loss": 0.4685714542865753,
      "eval_runtime": 25.5337,
      "eval_samples_per_second": 1583.399,
      "eval_steps_per_second": 6.188,
      "step": 7110
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.4561879634857178,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.146,
      "step": 8532
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8407370764283948,
      "eval_combined_score": 0.8117081041799742,
      "eval_f1": 0.7826791319315536,
      "eval_loss": 0.502971351146698,
      "eval_runtime": 25.5865,
      "eval_samples_per_second": 1580.13,
      "eval_steps_per_second": 6.175,
      "step": 8532
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.5375120639801025,
      "learning_rate": 4.3e-05,
      "loss": 0.1182,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8410586198367549,
      "eval_combined_score": 0.8122745929372454,
      "eval_f1": 0.7834905660377358,
      "eval_loss": 0.540908932685852,
      "eval_runtime": 25.5371,
      "eval_samples_per_second": 1583.189,
      "eval_steps_per_second": 6.187,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "step": 9954,
      "total_flos": 3.3506166766980096e+17,
      "train_loss": 0.25822540093119756,
      "train_runtime": 4293.4874,
      "train_samples_per_second": 4237.185,
      "train_steps_per_second": 16.56
    }
  ],
  "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
}