trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.711735188961029,
  "best_model_checkpoint": "bert_base_rand_20_v2_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": 2.1896302700042725,
      "learning_rate": 4.9e-05,
      "loss": 0.7328,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.711735188961029,
      "eval_runtime": 0.0563,
      "eval_samples_per_second": 1260.601,
      "eval_steps_per_second": 17.755,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.27385413646698,
      "learning_rate": 4.8e-05,
      "loss": 0.7151,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.323943661971831,
      "eval_loss": 0.717044472694397,
      "eval_runtime": 0.0562,
      "eval_samples_per_second": 1262.348,
      "eval_steps_per_second": 17.78,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.328735828399658,
      "learning_rate": 4.7e-05,
      "loss": 0.7065,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7503302097320557,
      "eval_runtime": 0.0582,
      "eval_samples_per_second": 1220.089,
      "eval_steps_per_second": 17.184,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.6860625743865967,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.7009,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5492957746478874,
      "eval_loss": 0.7191351652145386,
      "eval_runtime": 0.0549,
      "eval_samples_per_second": 1292.544,
      "eval_steps_per_second": 18.205,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.2746615409851074,
      "learning_rate": 4.5e-05,
      "loss": 0.698,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7310189008712769,
      "eval_runtime": 0.0757,
      "eval_samples_per_second": 938.37,
      "eval_steps_per_second": 13.216,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.7864902019500732,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6952,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.7655702233314514,
      "eval_runtime": 0.0667,
      "eval_samples_per_second": 1065.161,
      "eval_steps_per_second": 15.002,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.7080998553170098,
      "train_runtime": 19.2495,
      "train_samples_per_second": 1649.392,
      "train_steps_per_second": 7.792
    }
  ],
  "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
}