trainer_state.json

{
  "best_metric": 0.6933382749557495,
  "best_model_checkpoint": "bert_tiny_lda_100_v1_qnli/checkpoint-410",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 2460,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.14504414796829224,
      "learning_rate": 0.00098,
      "loss": 0.6982,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 1.6586,
      "eval_samples_per_second": 3293.755,
      "eval_steps_per_second": 13.264,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.21841265261173248,
      "learning_rate": 0.00096,
      "loss": 0.6931,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6936336159706116,
      "eval_runtime": 1.6485,
      "eval_samples_per_second": 3313.97,
      "eval_steps_per_second": 13.346,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.12709185481071472,
      "learning_rate": 0.00094,
      "loss": 0.6932,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 1.638,
      "eval_samples_per_second": 3335.131,
      "eval_steps_per_second": 13.431,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.024518601596355438,
      "learning_rate": 0.00092,
      "loss": 0.6932,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 1.6455,
      "eval_samples_per_second": 3319.898,
      "eval_steps_per_second": 13.37,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.047646842896938324,
      "learning_rate": 0.0009000000000000001,
      "loss": 0.6931,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4946000366099213,
      "eval_loss": 0.6934648156166077,
      "eval_runtime": 1.6442,
      "eval_samples_per_second": 3322.653,
      "eval_steps_per_second": 13.381,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.05455218255519867,
      "learning_rate": 0.00088,
      "loss": 0.6932,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5053999633900788,
      "eval_loss": 0.6933382749557495,
      "eval_runtime": 1.6522,
      "eval_samples_per_second": 3306.479,
      "eval_steps_per_second": 13.315,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "step": 2460,
      "total_flos": 1.6480417453111296e+16,
      "train_loss": 0.6940060127072218,
      "train_runtime": 346.8815,
      "train_samples_per_second": 15097.809,
      "train_steps_per_second": 59.098
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 1.6480417453111296e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}