in distilvit/curate.py [0:0]
def transform(self, captions, sentids):
transformed_captions = []
for caption, sentid in zip(captions, sentids):
result = self.transform_one(caption)
try:
grade = dict(
readability.getmeasures(result, lang="en")["readability grades"]
)
except Exception as e:
grade = {"DaleChallIndex": 10.0}
print(f"{caption} -> {result} with {grade['DaleChallIndex']:.2f}")
transformed_captions.append((result, grade, sentid))
transformed_captions.sort(key=self.by_grade)
return list(zip(*transformed_captions))