fix: save train/test.jsonl to output_dir not parent

This commit is contained in:
Christian Medina
2026-06-30 20:01:05 -04:00
parent d5ba471b98
commit 7181acc580

View File

@@ -62,11 +62,11 @@ def prepare_dataset(input_file, output_file, test_size=0.05):
test_data = formatted[split_point:] test_data = formatted[split_point:]
# Save # Save
with open(output_file.parent / "train.jsonl", "w") as f: with open(output_dir / "train.jsonl", "w") as f:
for item in train_data: for item in train_data:
f.write(json.dumps(item) + "\n") f.write(json.dumps(item) + "\n")
with open(output_file.parent / "test.jsonl", "w") as f: with open(output_dir / "test.jsonl", "w") as f:
for item in test_data: for item in test_data:
f.write(json.dumps(item) + "\n") f.write(json.dumps(item) + "\n")