trainer_state.json

{
  "best_global_step": 264,
  "best_metric": 0.4408361315727234,
  "best_model_checkpoint": "bert_base_km_5_v2_sst2/checkpoint-264",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 1584,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 6.928033828735352,
      "learning_rate": 4.9e-05,
      "loss": 0.3855,
      "step": 264
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8107798165137615,
      "eval_loss": 0.4408361315727234,
      "eval_runtime": 0.5801,
      "eval_samples_per_second": 1503.097,
      "eval_steps_per_second": 6.895,
      "step": 264
    },
    {
      "epoch": 2.0,
      "grad_norm": 11.186700820922852,
      "learning_rate": 4.8e-05,
      "loss": 0.2196,
      "step": 528
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8096330275229358,
      "eval_loss": 0.5447783470153809,
      "eval_runtime": 0.5683,
      "eval_samples_per_second": 1534.322,
      "eval_steps_per_second": 7.038,
      "step": 528
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.588103771209717,
      "learning_rate": 4.7e-05,
      "loss": 0.1531,
      "step": 792
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7947247706422018,
      "eval_loss": 0.5998702645301819,
      "eval_runtime": 0.5652,
      "eval_samples_per_second": 1542.937,
      "eval_steps_per_second": 7.078,
      "step": 792
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.448475360870361,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.1094,
      "step": 1056
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8061926605504587,
      "eval_loss": 0.6511033773422241,
      "eval_runtime": 0.5777,
      "eval_samples_per_second": 1509.389,
      "eval_steps_per_second": 6.924,
      "step": 1056
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.719103813171387,
      "learning_rate": 4.5e-05,
      "loss": 0.0831,
      "step": 1320
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7970183486238532,
      "eval_loss": 0.6421980261802673,
      "eval_runtime": 0.562,
      "eval_samples_per_second": 1551.736,
      "eval_steps_per_second": 7.118,
      "step": 1320
    },
    {
      "epoch": 6.0,
      "grad_norm": 13.903855323791504,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.0646,
      "step": 1584
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8038990825688074,
      "eval_loss": 0.8283168077468872,
      "eval_runtime": 0.5707,
      "eval_samples_per_second": 1527.917,
      "eval_steps_per_second": 7.009,
      "step": 1584
    },
    {
      "epoch": 6.0,
      "step": 1584,
      "total_flos": 5.316079940232192e+16,
      "train_loss": 0.1692113418771763,
      "train_runtime": 662.1963,
      "train_samples_per_second": 5085.274,
      "train_steps_per_second": 19.934
    }
  ],
  "logging_steps": 1,
  "max_steps": 13200,
  "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": 5.316079940232192e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}