Update NCTCRCHE100K.py

NCTCRCHE100K.py

@@ -61,8 +61,10 @@ class NCTCRCHE100K(datasets.GeneratorBasedBuilder):
         idx = 0
         for filepath, image in images:
             # extract text from each item
+            start_index = filepath.find('/') + 1
+            end_index = filepath.find('-', start_index)
             yield idx, {
                 "image": {"path": filepath, "bytes": image.read()},
-                "text": filepath
+                "text": filepath[start_index: end_index]
             }
             idx += 1