artifacts/model/bundle/data_config.json

{
  "dataloader": {
    "batch_size": 16,
    "num_workers": 0,
    "pin_memory": false
  },
  "dataset": {
    "config": null,
    "debug_subset_size": 1024,
    "label_field": "label",
    "name": "ag_news",
    "random_seed": 42,
    "text_field": "text",
    "validation_split_ratio": 0.1
  },
  "paths": {
    "dataset_cache_dir": "data\\hf_cache",
    "processed_cache_dir": "data\\processed",
    "tokenizer_dir": "artifacts\\tokenizer"
  },
  "tokenizer": {
    "max_sequence_length": 256,
    "min_frequency": 2,
    "vocab_size": 16000
  }
}