Update handwriting_dataset.py

handwriting_dataset.py

@@ -44,7 +44,7 @@ class HandwritingData(datasets.GeneratorBasedBuilder):
         # Iterate through images
         for idx, filepath, image in enumerate(images):
             # extract the text from the filename
-            text = filepath.split("/")[-1].split(".")[0]
+            text = [c for c in filepath if not 0 <= ord(c) <= 127][0]
             yield idx, {
                 "text": text,
                 "image": {"path": filepath, "bytes": image.read()}