checkpoint-4000/trainer_state.json

{
  "best_metric": 6.433775901794434,
  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained_last/de_mlm/de_childes_42/checkpoint-4000",
  "epoch": 3.0041306796845664,
  "eval_steps": 2000,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.5020653398422832,
      "eval_loss": 7.505205154418945,
      "eval_runtime": 4.9012,
      "eval_samples_per_second": 677.383,
      "eval_steps_per_second": 42.438,
      "step": 2000
    },
    {
      "epoch": 3.0041306796845664,
      "grad_norm": 1.3569484949111938,
      "learning_rate": 1e-05,
      "loss": 7.4063,
      "step": 4000
    },
    {
      "epoch": 3.0041306796845664,
      "eval_loss": 6.433775901794434,
      "eval_runtime": 4.7916,
      "eval_samples_per_second": 692.873,
      "eval_steps_per_second": 43.409,
      "step": 4000
    }
  ],
  "logging_steps": 4000,
  "max_steps": 100000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 76,
  "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": 1247512767234048.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}