checkpoint-4716/trainer_state.json

{
  "best_metric": 0.7884615384615384,
  "best_model_checkpoint": "/home/ubuntu/utah/babylm-24/src/evaluation/results/finetune/DebertaV2-Base-10M_babylm-A/boolq/checkpoint-4716",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 4716,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.42408821034775235,
      "grad_norm": 4.898578643798828,
      "learning_rate": 9.787955894826124e-06,
      "loss": 0.669,
      "step": 500
    },
    {
      "epoch": 0.8481764206955047,
      "grad_norm": 2.427595615386963,
      "learning_rate": 9.575911789652248e-06,
      "loss": 0.6535,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6513761281967163,
      "eval_f1": 0.7812739831158864,
      "eval_loss": 0.6350241899490356,
      "eval_mcc": 0.1190648771022687,
      "eval_runtime": 2.7173,
      "eval_samples_per_second": 601.699,
      "eval_steps_per_second": 75.442,
      "step": 1179
    },
    {
      "epoch": 1.272264631043257,
      "grad_norm": 3.5514039993286133,
      "learning_rate": 9.363867684478373e-06,
      "loss": 0.6279,
      "step": 1500
    },
    {
      "epoch": 1.6963528413910094,
      "grad_norm": 6.262450218200684,
      "learning_rate": 9.151823579304497e-06,
      "loss": 0.6019,
      "step": 2000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6446483135223389,
      "eval_f1": 0.7172749391727494,
      "eval_loss": 0.6267324090003967,
      "eval_mcc": 0.23986158954968526,
      "eval_runtime": 2.7281,
      "eval_samples_per_second": 599.313,
      "eval_steps_per_second": 75.143,
      "step": 2358
    },
    {
      "epoch": 2.1204410517387617,
      "grad_norm": 10.400861740112305,
      "learning_rate": 8.93977947413062e-06,
      "loss": 0.5841,
      "step": 2500
    },
    {
      "epoch": 2.544529262086514,
      "grad_norm": 6.1930718421936035,
      "learning_rate": 8.727735368956744e-06,
      "loss": 0.5239,
      "step": 3000
    },
    {
      "epoch": 2.9686174724342664,
      "grad_norm": 9.404278755187988,
      "learning_rate": 8.515691263782867e-06,
      "loss": 0.5263,
      "step": 3500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6709480285644531,
      "eval_f1": 0.738581146744412,
      "eval_loss": 0.6594865322113037,
      "eval_mcc": 0.2953818504855689,
      "eval_runtime": 2.7221,
      "eval_samples_per_second": 600.629,
      "eval_steps_per_second": 75.308,
      "step": 3537
    },
    {
      "epoch": 3.392705682782019,
      "grad_norm": 8.5890474319458,
      "learning_rate": 8.303647158608991e-06,
      "loss": 0.4231,
      "step": 4000
    },
    {
      "epoch": 3.816793893129771,
      "grad_norm": 19.422046661376953,
      "learning_rate": 8.091603053435115e-06,
      "loss": 0.4158,
      "step": 4500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7039755582809448,
      "eval_f1": 0.7884615384615384,
      "eval_loss": 0.7751796245574951,
      "eval_mcc": 0.3197717100660696,
      "eval_runtime": 2.7784,
      "eval_samples_per_second": 588.466,
      "eval_steps_per_second": 73.783,
      "step": 4716
    }
  ],
  "logging_steps": 500,
  "max_steps": 23580,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 3,
        "early_stopping_threshold": 0.001
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2224686723207168.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}