Update layoutlmv3.py

layoutlmv3.py

@@ -63,6 +63,7 @@ class Funsd(datasets.GeneratorBasedBuilder):
                 {
                     "id": datasets.Value("string"),
                     "tokens": datasets.Sequence(datasets.Value("string")),
+                    "words": datasets.Sequence(datasets.Value("string")),
                     "bboxes": datasets.Sequence(datasets.Sequence(datasets.Value("int64"))),
                     "ner_tags": datasets.Sequence(
                         datasets.features.ClassLabel(
@@ -106,6 +107,7 @@ class Funsd(datasets.GeneratorBasedBuilder):
         img_dir = os.path.join(filepath, "images")
         for guid, file in enumerate(sorted(os.listdir(ann_dir))):
             tokens = []
+            words = []
             bboxes = []
             ner_tags = []
 
@@ -121,6 +123,7 @@ class Funsd(datasets.GeneratorBasedBuilder):
                         word=item['text']
                         ner_tags.append(labels)
                         tokens.append(word)
+                        words.append(word)
                         bboxes.append(normalize_bbox(item['box'],size))
                 
                     #cur_line_bboxes = []
@@ -143,5 +146,5 @@ class Funsd(datasets.GeneratorBasedBuilder):
                             #cur_line_bboxes.append(normalize_bbox(w["box"], size))
                     #cur_line_bboxes = self.get_line_bbox(cur_line_bboxes)
                     #bboxes.extend(cur_line_bboxes)
-            yield guid, {"id": str(guid), "tokens": tokens, "bboxes": bboxes, "ner_tags": ner_tags,
+            yield guid, {"id": str(guid), "tokens": tokens,"words":words, "bboxes": bboxes, "ner_tags": ner_tags,
                          "image": image, "image_path":image_path}