MrPotato commited on
Commit
c9a9db0
·
1 Parent(s): 8678d26
Files changed (1) hide show
  1. ref_seg_ger.py +4 -1
ref_seg_ger.py CHANGED
@@ -235,13 +235,16 @@ class RefSeg(datasets.GeneratorBasedBuilder):
235
  for input_id in tokenized_input['input_ids']:
236
  input_ids.append(input_id)
237
  labels.append('O')
238
- else:
239
  input_ids.append(tokenized_input['input_ids'])
240
  if row['tag'] == 'O':
241
  labels.append(row['tag'])
242
  else:
243
  labels.append(row['tag'] + '-' + row['label'])
 
 
244
 
 
245
  for chunk_id, index in enumerate(range(0, len(input_ids), self.CHUNK_SIZE)):
246
  split_ids = input_ids[index:index + self.CHUNK_SIZE]
247
  #split_bboxes = bboxes[index:index + self.CHUNK_SIZE]
 
235
  for input_id in tokenized_input['input_ids']:
236
  input_ids.append(input_id)
237
  labels.append('O')
238
+ elif len(tokenized_input['input_ids']) == 1:
239
  input_ids.append(tokenized_input['input_ids'])
240
  if row['tag'] == 'O':
241
  labels.append(row['tag'])
242
  else:
243
  labels.append(row['tag'] + '-' + row['label'])
244
+ else:
245
+ continue
246
 
247
+ print('got all')
248
  for chunk_id, index in enumerate(range(0, len(input_ids), self.CHUNK_SIZE)):
249
  split_ids = input_ids[index:index + self.CHUNK_SIZE]
250
  #split_bboxes = bboxes[index:index + self.CHUNK_SIZE]