0 indexing the labels
Browse files- create_dataset.py +2 -1
create_dataset.py
CHANGED
|
@@ -6,11 +6,12 @@ def main():
|
|
| 6 |
raw_data = load_dataset("amazon_reviews_multi", "ja")
|
| 7 |
raw_data = raw_data.rename_column("review_id", "id")
|
| 8 |
raw_data = raw_data.rename_column("review_body", "text")
|
| 9 |
-
raw_data = raw_data.rename_column("stars", "label")
|
| 10 |
raw_data = raw_data.remove_columns(["product_id", "reviewer_id", "review_title", "language", "product_category"])
|
| 11 |
|
| 12 |
for split, dataset in raw_data.items():
|
|
|
|
| 13 |
dataset = dataset.map(lambda x: {"label_text": str(x["label"])}, num_proc=4)
|
|
|
|
| 14 |
dataset.to_json(f"{split}.jsonl")
|
| 15 |
|
| 16 |
if __name__ == "__main__":
|
|
|
|
| 6 |
raw_data = load_dataset("amazon_reviews_multi", "ja")
|
| 7 |
raw_data = raw_data.rename_column("review_id", "id")
|
| 8 |
raw_data = raw_data.rename_column("review_body", "text")
|
|
|
|
| 9 |
raw_data = raw_data.remove_columns(["product_id", "reviewer_id", "review_title", "language", "product_category"])
|
| 10 |
|
| 11 |
for split, dataset in raw_data.items():
|
| 12 |
+
dataset = dataset.map(lambda x: {"label": x["stars"]-1}, num_proc=4)
|
| 13 |
dataset = dataset.map(lambda x: {"label_text": str(x["label"])}, num_proc=4)
|
| 14 |
+
dataset = dataset.remove_columns(["stars"])
|
| 15 |
dataset.to_json(f"{split}.jsonl")
|
| 16 |
|
| 17 |
if __name__ == "__main__":
|