checkpoint-4000/trainer_state.json

{
  "best_metric": 7.107393264770508,
  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained_last/en_wiki_mlm_30/checkpoint-4000",
  "epoch": 2.2637238256932655,
  "eval_steps": 2000,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.1318619128466327,
      "eval_loss": 7.872045516967773,
      "eval_runtime": 3.9374,
      "eval_samples_per_second": 1252.08,
      "eval_steps_per_second": 78.477,
      "step": 2000
    },
    {
      "epoch": 2.2637238256932655,
      "grad_norm": 1.0014455318450928,
      "learning_rate": 1e-05,
      "loss": 7.9236,
      "step": 4000
    },
    {
      "epoch": 2.2637238256932655,
      "eval_loss": 7.107393264770508,
      "eval_runtime": 3.9376,
      "eval_samples_per_second": 1252.036,
      "eval_steps_per_second": 78.474,
      "step": 4000
    }
  ],
  "logging_steps": 4000,
  "max_steps": 100000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 57,
  "save_steps": 4000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1247629743489024.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}