batched tokenization

docbank.py

@@ -60,7 +60,7 @@ _URLS = {
 _FEATURES = datasets.Features(
     {
         "id": datasets.Value("string"),
-        "input_ids": datasets.Sequence(datasets.Value("int64")),
+        "input_ids": datasets.Sequence(datasets.Value("string")),
         "bbox": datasets.Sequence(datasets.Sequence(datasets.Value("int64"))),
         # "RGBs": datasets.Sequence(datasets.Sequence(datasets.Value("int64"))),
         # "fonts": datasets.Sequence(datasets.Value("string")),
@@ -267,42 +267,29 @@ class Docbank(datasets.GeneratorBasedBuilder):
                 continue
 
             #print('Processing...')
-            processed = self.TOKENIZER(
-                tokens,
-                boxes=bboxes,
-                word_labels=labels,
-                padding=True,
-                max_length=512,
-                truncation=True,
-                return_overflowing_tokens=True,
-                return_offsets_mapping=True
-
-            )
+            # processed = self.TOKENIZER(
+            #     tokens,
+            #     boxes=bboxes,
+            #     word_labels=labels,
+            #     add_special_tokens=False,
+            #     return_offsets_mapping=False,
+            #     return_attention_mask=False,
+            # )
             #print(processed)
 
-            for batch_id, batch in enumerate(processed['input_ids']):
-                split_tokens = batch
-                split_bboxes = processed['bbox'][batch_id]
-                split_attention_mask = processed['attention_mask'][batch_id]
-                split_offset_mapping = processed['offset_mapping'][batch_id]
+            for chunk_id, index in enumerate(range(0, len(tokens), self.CHUNK_SIZE)):
+                split_tokens = tokens[index:index + self.CHUNK_SIZE]
+                split_bboxes = bboxes[index:index + self.CHUNK_SIZE]
                 # split_rgbs = rgbs[index:index + self.CHUNK_SIZE]
                 # split_fonts = fonts[index:index + self.CHUNK_SIZE]
-                split_labels = processed['labels'][batch_id]
-            # for chunk_id, index in enumerate(range(0, len(processed['input_ids']), self.CHUNK_SIZE)):
-            #     split_tokens = processed['input_ids'][index:index + self.CHUNK_SIZE]
-            #     split_bboxes = processed['bbox'][index:index + self.CHUNK_SIZE]
-            #     # split_rgbs = rgbs[index:index + self.CHUNK_SIZE]
-            #     # split_fonts = fonts[index:index + self.CHUNK_SIZE]
-            #     split_labels = processed['labels'][index:index + self.CHUNK_SIZE]
+                split_labels = labels[index:index + self.CHUNK_SIZE]
 
                 #tokenized = self.TOKENIZER(processed['words'], boxes=processed['boxes'])
 
                 yield key, {
-                    "id": f"{f_id}_{batch_id}",
+                    "id": f"{f_id}_{chunk_id}",
                     'input_ids': split_tokens,
                     "bbox": split_bboxes,
-                    "attention_mask": split_attention_mask,
-                    "offset_mapping":split_offset_mapping,
                     # "RGBs": split_rgbs,
                     # "fonts": split_fonts,
                     #"image": image,