trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5970614552497864,
  "best_model_checkpoint": "distilbert_km_10_v1_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": 1.3870301246643066,
      "learning_rate": 4.9e-05,
      "loss": 0.6306,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6862745098039216,
      "eval_combined_score": 0.739241151005857,
      "eval_f1": 0.7922077922077922,
      "eval_loss": 0.6149829626083374,
      "eval_runtime": 0.1891,
      "eval_samples_per_second": 2157.275,
      "eval_steps_per_second": 10.575,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.859695553779602,
      "learning_rate": 4.8e-05,
      "loss": 0.5631,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6985294117647058,
      "eval_combined_score": 0.7556573999462799,
      "eval_f1": 0.8127853881278538,
      "eval_loss": 0.5970614552497864,
      "eval_runtime": 0.183,
      "eval_samples_per_second": 2229.673,
      "eval_steps_per_second": 10.93,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.6454362869262695,
      "learning_rate": 4.7e-05,
      "loss": 0.485,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7009803921568627,
      "eval_combined_score": 0.7530461066215624,
      "eval_f1": 0.805111821086262,
      "eval_loss": 0.6306254267692566,
      "eval_runtime": 0.2027,
      "eval_samples_per_second": 2012.52,
      "eval_steps_per_second": 9.865,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.5285091400146484,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.3824,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6617647058823529,
      "eval_combined_score": 0.7115051903114187,
      "eval_f1": 0.7612456747404844,
      "eval_loss": 0.7183700203895569,
      "eval_runtime": 0.1917,
      "eval_samples_per_second": 2128.665,
      "eval_steps_per_second": 10.435,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.753702402114868,
      "learning_rate": 4.5e-05,
      "loss": 0.2587,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6200980392156863,
      "eval_combined_score": 0.6640979838262988,
      "eval_f1": 0.7080979284369114,
      "eval_loss": 0.9183156490325928,
      "eval_runtime": 0.1901,
      "eval_samples_per_second": 2146.142,
      "eval_steps_per_second": 10.52,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.3432817459106445,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.152,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6225490196078431,
      "eval_combined_score": 0.6637649312598602,
      "eval_f1": 0.7049808429118773,
      "eval_loss": 1.0391442775726318,
      "eval_runtime": 0.2245,
      "eval_samples_per_second": 1817.496,
      "eval_steps_per_second": 8.909,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.381226539611816,
      "learning_rate": 4.3e-05,
      "loss": 0.0757,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6225490196078431,
      "eval_combined_score": 0.6702488687782806,
      "eval_f1": 0.717948717948718,
      "eval_loss": 1.3927804231643677,
      "eval_runtime": 0.2063,
      "eval_samples_per_second": 1977.828,
      "eval_steps_per_second": 9.695,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 1700616463945728.0,
      "train_loss": 0.36394872665405276,
      "train_runtime": 34.4408,
      "train_samples_per_second": 5325.08,
      "train_steps_per_second": 21.776
    }
  ],
  "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
}