trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7413071393966675,
  "best_model_checkpoint": "bert_base_km_100_v1_wnli/checkpoint-3",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.6636451482772827,
      "learning_rate": 4.9e-05,
      "loss": 0.7453,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.36619718309859156,
      "eval_loss": 0.7413071393966675,
      "eval_runtime": 0.0557,
      "eval_samples_per_second": 1273.627,
      "eval_steps_per_second": 17.938,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2630294561386108,
      "learning_rate": 4.8e-05,
      "loss": 0.6965,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.22535211267605634,
      "eval_loss": 0.7976452708244324,
      "eval_runtime": 0.0544,
      "eval_samples_per_second": 1304.296,
      "eval_steps_per_second": 18.37,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.2269364595413208,
      "learning_rate": 4.7e-05,
      "loss": 0.6931,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.2535211267605634,
      "eval_loss": 0.811454713344574,
      "eval_runtime": 0.0556,
      "eval_samples_per_second": 1277.331,
      "eval_steps_per_second": 17.991,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.809952735900879,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6928,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.2676056338028169,
      "eval_loss": 0.8429247736930847,
      "eval_runtime": 0.0558,
      "eval_samples_per_second": 1271.946,
      "eval_steps_per_second": 17.915,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.1863123178482056,
      "learning_rate": 4.5e-05,
      "loss": 0.6768,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.16901408450704225,
      "eval_loss": 0.894283652305603,
      "eval_runtime": 0.0594,
      "eval_samples_per_second": 1195.126,
      "eval_steps_per_second": 16.833,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.439569354057312,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6739,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.1267605633802817,
      "eval_loss": 0.9332910180091858,
      "eval_runtime": 0.0582,
      "eval_samples_per_second": 1220.829,
      "eval_steps_per_second": 17.195,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.6964046955108643,
      "train_runtime": 18.1355,
      "train_samples_per_second": 1750.711,
      "train_steps_per_second": 8.271
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "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": 501226560460800.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}