trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.5897228121757507,
  "best_model_checkpoint": "bert_base_rand_50_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": 0.9206721782684326,
      "learning_rate": 4.9e-05,
      "loss": 0.6271,
      "step": 15
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6911764705882353,
      "eval_combined_score": 0.7459046909903202,
      "eval_f1": 0.8006329113924051,
      "eval_loss": 0.5931947231292725,
      "eval_runtime": 0.2806,
      "eval_samples_per_second": 1454.107,
      "eval_steps_per_second": 7.128,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.3034330606460571,
      "learning_rate": 4.8e-05,
      "loss": 0.5729,
      "step": 30
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7058823529411765,
      "eval_combined_score": 0.7494929006085194,
      "eval_f1": 0.7931034482758621,
      "eval_loss": 0.5897228121757507,
      "eval_runtime": 0.3228,
      "eval_samples_per_second": 1263.903,
      "eval_steps_per_second": 6.196,
      "step": 30
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.0921132564544678,
      "learning_rate": 4.7e-05,
      "loss": 0.4928,
      "step": 45
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6862745098039216,
      "eval_combined_score": 0.7271952259164536,
      "eval_f1": 0.7681159420289855,
      "eval_loss": 0.6286938190460205,
      "eval_runtime": 0.296,
      "eval_samples_per_second": 1378.552,
      "eval_steps_per_second": 6.758,
      "step": 45
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.4250071048736572,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.3613,
      "step": 60
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.678921568627451,
      "eval_combined_score": 0.7210159380207779,
      "eval_f1": 0.7631103074141049,
      "eval_loss": 0.739713191986084,
      "eval_runtime": 0.2829,
      "eval_samples_per_second": 1441.963,
      "eval_steps_per_second": 7.068,
      "step": 60
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.448269367218018,
      "learning_rate": 4.5e-05,
      "loss": 0.2399,
      "step": 75
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6593137254901961,
      "eval_combined_score": 0.7007143766597548,
      "eval_f1": 0.7421150278293135,
      "eval_loss": 0.9837620854377747,
      "eval_runtime": 0.2821,
      "eval_samples_per_second": 1446.335,
      "eval_steps_per_second": 7.09,
      "step": 75
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.104319095611572,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1438,
      "step": 90
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6225490196078431,
      "eval_combined_score": 0.6578880556206547,
      "eval_f1": 0.6932270916334662,
      "eval_loss": 1.201783299446106,
      "eval_runtime": 0.278,
      "eval_samples_per_second": 1467.65,
      "eval_steps_per_second": 7.194,
      "step": 90
    },
    {
      "epoch": 7.0,
      "grad_norm": 8.028135299682617,
      "learning_rate": 4.3e-05,
      "loss": 0.099,
      "step": 105
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6299019607843137,
      "eval_combined_score": 0.6694789187351241,
      "eval_f1": 0.7090558766859345,
      "eval_loss": 1.4124648571014404,
      "eval_runtime": 0.2719,
      "eval_samples_per_second": 1500.553,
      "eval_steps_per_second": 7.356,
      "step": 105
    },
    {
      "epoch": 7.0,
      "step": 105,
      "total_flos": 3377819728711680.0,
      "train_loss": 0.362400636218843,
      "train_runtime": 59.1657,
      "train_samples_per_second": 3099.77,
      "train_steps_per_second": 12.676
    }
  ],
  "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": 3377819728711680.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}