| from datasets import load_dataset | |
| def main(): | |
| raw_dset = load_dataset("ag_news") | |
| labels = raw_dset["train"].features["label"] | |
| for split, dset in raw_dset.items(): | |
| dset = dset.map(lambda x: {"label_text": labels.int2str(x["label"])}, num_proc=4) | |
| dset.to_json(f"{split}.jsonl") | |
| if __name__ == "__main__": | |
| main() | |