Download create_dataset.py from SetFit/mrpc: direct link, hf CLI and curl.
- Browser
- Download file 685 Bytes
-
https://huggingface.co/datasets/SetFit/mrpc/resolve/main/create_dataset.py
- Command line
-
hf download hf://datasets/SetFit/mrpc/create_dataset.py
-
curl -L -o create_dataset.py https://huggingface.co/datasets/SetFit/mrpc/resolve/main/create_dataset.py
685 Bytes
| from datasets import load_dataset | |
| def main(): | |
| raw_dset = load_dataset("glue", "mrpc") | |
| raw_dset = raw_dset.rename_column("sentence1", "text1") | |
| raw_dset = raw_dset.rename_column("sentence2", "text2") | |
| text_labs = ["not equivalent", "equivalent"] | |
| for split, dset in raw_dset.items(): | |
| label_text = [] | |
| for i in dset["label"]: | |
| if i >= 0: | |
| label_text.append(text_labs[i]) | |
| else: | |
| label_text.append("unlabeled") | |
| dset = dset.add_column("label_text", label_text) | |
| dset.to_json(f"{split}.jsonl") | |
| if __name__ == "__main__": | |
| main() | |
| print("Job's done!") |