trainer_state.json

{
  "best_global_step": 20,
  "best_metric": 0.6887127757072449,
  "best_model_checkpoint": "bert_base_km_5_v2_rte/checkpoint-20",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 70,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.3064942359924316,
      "learning_rate": 4.9e-05,
      "loss": 0.7148,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4657039711191336,
      "eval_loss": 0.705635130405426,
      "eval_runtime": 0.1914,
      "eval_samples_per_second": 1447.317,
      "eval_steps_per_second": 10.45,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.913057804107666,
      "learning_rate": 4.8e-05,
      "loss": 0.6543,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5703971119133574,
      "eval_loss": 0.6887127757072449,
      "eval_runtime": 0.1992,
      "eval_samples_per_second": 1390.804,
      "eval_steps_per_second": 10.042,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.5929436683654785,
      "learning_rate": 4.7e-05,
      "loss": 0.588,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.48375451263537905,
      "eval_loss": 0.7249900698661804,
      "eval_runtime": 0.1896,
      "eval_samples_per_second": 1460.882,
      "eval_steps_per_second": 10.548,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.6172940731048584,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4714,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5415162454873647,
      "eval_loss": 0.8258281350135803,
      "eval_runtime": 0.1902,
      "eval_samples_per_second": 1456.003,
      "eval_steps_per_second": 10.513,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.299875020980835,
      "learning_rate": 4.5e-05,
      "loss": 0.3136,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5523465703971119,
      "eval_loss": 1.0116863250732422,
      "eval_runtime": 0.1889,
      "eval_samples_per_second": 1466.479,
      "eval_steps_per_second": 10.588,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.938107967376709,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2059,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5379061371841155,
      "eval_loss": 1.319309115409851,
      "eval_runtime": 0.1913,
      "eval_samples_per_second": 1448.181,
      "eval_steps_per_second": 10.456,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.1681761741638184,
      "learning_rate": 4.3e-05,
      "loss": 0.1215,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5342960288808665,
      "eval_loss": 1.473266839981079,
      "eval_runtime": 0.1925,
      "eval_samples_per_second": 1439.263,
      "eval_steps_per_second": 10.392,
      "step": 70
    },
    {
      "epoch": 7.0,
      "step": 70,
      "total_flos": 2293012847462400.0,
      "train_loss": 0.43848816667284285,
      "train_runtime": 44.1953,
      "train_samples_per_second": 2817.043,
      "train_steps_per_second": 11.313
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "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": 2293012847462400.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}