trainer_state.json

{
  "best_metric": 0.31332626938819885,
  "best_model_checkpoint": "distilbert_lda_100_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": 2.956484794616699,
      "learning_rate": 4.9e-05,
      "loss": 0.4043,
      "step": 1422
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8511006678209251,
      "eval_combined_score": 0.8251500670658996,
      "eval_f1": 0.7991994663108739,
      "eval_loss": 0.32754120230674744,
      "eval_runtime": 17.107,
      "eval_samples_per_second": 2363.359,
      "eval_steps_per_second": 9.236,
      "step": 1422
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.963893413543701,
      "learning_rate": 4.8e-05,
      "loss": 0.2918,
      "step": 2844
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8599554786049963,
      "eval_combined_score": 0.8419983745191059,
      "eval_f1": 0.8240412704332154,
      "eval_loss": 0.31332626938819885,
      "eval_runtime": 17.2399,
      "eval_samples_per_second": 2345.134,
      "eval_steps_per_second": 9.165,
      "step": 2844
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.0235133171081543,
      "learning_rate": 4.7e-05,
      "loss": 0.2305,
      "step": 4266
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8714568389809547,
      "eval_combined_score": 0.852738415977335,
      "eval_f1": 0.8340199929737153,
      "eval_loss": 0.3146589696407318,
      "eval_runtime": 17.2803,
      "eval_samples_per_second": 2339.665,
      "eval_steps_per_second": 9.143,
      "step": 4266
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.2095916271209717,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.179,
      "step": 5688
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8760079149146673,
      "eval_combined_score": 0.8519556894083102,
      "eval_f1": 0.8279034639019532,
      "eval_loss": 0.31782716512680054,
      "eval_runtime": 17.1901,
      "eval_samples_per_second": 2351.932,
      "eval_steps_per_second": 9.191,
      "step": 5688
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.359029531478882,
      "learning_rate": 4.5e-05,
      "loss": 0.1389,
      "step": 7110
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8804600544150384,
      "eval_combined_score": 0.8585003961999769,
      "eval_f1": 0.8365407379849155,
      "eval_loss": 0.3524511456489563,
      "eval_runtime": 17.2563,
      "eval_samples_per_second": 2342.907,
      "eval_steps_per_second": 9.156,
      "step": 7110
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.7366631031036377,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1067,
      "step": 8532
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8782587187731882,
      "eval_combined_score": 0.8595780276048511,
      "eval_f1": 0.8408973364365141,
      "eval_loss": 0.3905220925807953,
      "eval_runtime": 17.273,
      "eval_samples_per_second": 2340.65,
      "eval_steps_per_second": 9.147,
      "step": 8532
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.230625629425049,
      "learning_rate": 4.3e-05,
      "loss": 0.086,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8788276032649023,
      "eval_combined_score": 0.8607857401295461,
      "eval_f1": 0.84274387699419,
      "eval_loss": 0.4037318527698517,
      "eval_runtime": 17.1931,
      "eval_samples_per_second": 2351.532,
      "eval_steps_per_second": 9.19,
      "step": 9954
    },
    {
      "epoch": 7.0,
      "step": 9954,
      "total_flos": 1.686920659598684e+17,
      "train_loss": 0.20531803780299535,
      "train_runtime": 2492.9036,
      "train_samples_per_second": 7297.635,
      "train_steps_per_second": 28.521
    }
  ],
  "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": 1.686920659598684e+17,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}