trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5786468386650085,
  "best_model_checkpoint": "distilbert_rand_10_v2_mrpc/checkpoint-30",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 105,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.2592852115631104,
      "learning_rate": 4.9e-05,
      "loss": 0.6329,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6862745098039216,
      "eval_combined_score": 0.7405731523378583,
      "eval_f1": 0.7948717948717948,
      "eval_loss": 0.5951622128486633,
      "eval_runtime": 0.1804,
      "eval_samples_per_second": 2261.454,
      "eval_steps_per_second": 11.086,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6964359283447266,
      "learning_rate": 4.8e-05,
      "loss": 0.5742,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6838235294117647,
      "eval_combined_score": 0.7353001944579485,
      "eval_f1": 0.7867768595041322,
      "eval_loss": 0.5786468386650085,
      "eval_runtime": 0.2075,
      "eval_samples_per_second": 1966.351,
      "eval_steps_per_second": 9.639,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.5742881298065186,
      "learning_rate": 4.7e-05,
      "loss": 0.5006,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6838235294117647,
      "eval_combined_score": 0.7370337159253946,
      "eval_f1": 0.7902439024390244,
      "eval_loss": 0.6243956089019775,
      "eval_runtime": 0.2003,
      "eval_samples_per_second": 2037.374,
      "eval_steps_per_second": 9.987,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.952268600463867,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.3971,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7491612924239464,
      "eval_f1": 0.7973421926910299,
      "eval_loss": 0.7714180946350098,
      "eval_runtime": 0.2098,
      "eval_samples_per_second": 1944.851,
      "eval_steps_per_second": 9.534,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.923924446105957,
      "learning_rate": 4.5e-05,
      "loss": 0.2599,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6642156862745098,
      "eval_combined_score": 0.7082380420522639,
      "eval_f1": 0.7522603978300181,
      "eval_loss": 0.9506104588508606,
      "eval_runtime": 0.2172,
      "eval_samples_per_second": 1878.407,
      "eval_steps_per_second": 9.208,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.3849189281463623,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1453,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6397058823529411,
      "eval_combined_score": 0.6834893048128342,
      "eval_f1": 0.7272727272727273,
      "eval_loss": 1.2577589750289917,
      "eval_runtime": 0.185,
      "eval_samples_per_second": 2205.778,
      "eval_steps_per_second": 10.813,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 5.978302001953125,
      "learning_rate": 4.3e-05,
      "loss": 0.0893,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6323529411764706,
      "eval_combined_score": 0.6783088235294117,
      "eval_f1": 0.7242647058823529,
      "eval_loss": 1.5317364931106567,
      "eval_runtime": 0.1809,
      "eval_samples_per_second": 2255.284,
      "eval_steps_per_second": 11.055,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.37131280444917225,
      "train_runtime": 34.4007,
      "train_samples_per_second": 5331.28,
      "train_steps_per_second": 21.802
    }
  ],
  "logging_steps": 1,
  "max_steps": 750,
  "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": 1700616463945728.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}