Upload docbank.py
Browse files- docbank.py +8 -0
docbank.py
CHANGED
|
@@ -247,12 +247,20 @@ class Docbank(datasets.GeneratorBasedBuilder):
|
|
| 247 |
continue
|
| 248 |
|
| 249 |
for chunk_id, index in enumerate(range(0, len(tokens), self.CHUNK_SIZE)):
|
|
|
|
| 250 |
split_tokens = tokens[index:index + self.CHUNK_SIZE]
|
| 251 |
split_bboxes = bboxes[index:index + self.CHUNK_SIZE]
|
| 252 |
split_rgbs = rgbs[index:index + self.CHUNK_SIZE]
|
| 253 |
split_fonts = fonts[index:index + self.CHUNK_SIZE]
|
| 254 |
split_labels = labels[index:index + self.CHUNK_SIZE]
|
| 255 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 256 |
yield key, {
|
| 257 |
"id": f"{f_id}_{chunk_id}",
|
| 258 |
'input_ids': split_tokens,
|
|
|
|
| 247 |
continue
|
| 248 |
|
| 249 |
for chunk_id, index in enumerate(range(0, len(tokens), self.CHUNK_SIZE)):
|
| 250 |
+
|
| 251 |
split_tokens = tokens[index:index + self.CHUNK_SIZE]
|
| 252 |
split_bboxes = bboxes[index:index + self.CHUNK_SIZE]
|
| 253 |
split_rgbs = rgbs[index:index + self.CHUNK_SIZE]
|
| 254 |
split_fonts = fonts[index:index + self.CHUNK_SIZE]
|
| 255 |
split_labels = labels[index:index + self.CHUNK_SIZE]
|
| 256 |
|
| 257 |
+
if len(split_tokens) > self.CHUNK_SIZE:
|
| 258 |
+
print('Err')
|
| 259 |
+
print(key)
|
| 260 |
+
print(f_id)
|
| 261 |
+
print(split_tokens)
|
| 262 |
+
|
| 263 |
+
|
| 264 |
yield key, {
|
| 265 |
"id": f"{f_id}_{chunk_id}",
|
| 266 |
'input_ids': split_tokens,
|