MrPotato commited on
Commit
c71cc00
·
1 Parent(s): b65b8c9

changed tokenization

Browse files
Files changed (1) hide show
  1. ref_seg_ger.py +5 -10
ref_seg_ger.py CHANGED
@@ -142,15 +142,10 @@ class RefSeg(datasets.GeneratorBasedBuilder):
142
  # ]
143
 
144
  # DEFAULT_CONFIG_NAME = "small" # It's not mandatory to have a default configuration. Just use one if it make sense.
145
- TOKENIZER = pre_tokenizer = pre_tokenizers.Sequence([
146
- Whitespace(),
147
- Digits(),
148
- Split(".", behavior="isolated"),
149
- Split(":", behavior="isolated"),
150
- Split("/", behavior="isolated"),
151
- Split("-", behavior="isolated"),
152
- Split(",", behavior="isolated"),
153
- ])
154
 
155
  def _info(self):
156
  # TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
@@ -285,7 +280,7 @@ class RefSeg(datasets.GeneratorBasedBuilder):
285
  split_ids = np.array_split(clean_input_ids, n_chunks)
286
  split_labels = np.array_split(clean_labels, n_chunks)
287
  split_refs = np.array_split(clean_refs, n_chunks)
288
- for chunk_ids, chunk_labels, chunk_refs in zip(split_ids, split_labels, split_refs):
289
  # for chunk_id, index in enumerate(range(0, len(clean_input_ids), self.CHUNK_SIZE)):
290
  # split_ids = clean_input_ids[index:max(len(clean_input_ids), index + self.CHUNK_SIZE)]
291
  # split_bboxes = bboxes[index:index + self.CHUNK_SIZE]
 
142
  # ]
143
 
144
  # DEFAULT_CONFIG_NAME = "small" # It's not mandatory to have a default configuration. Just use one if it make sense.
145
+
146
+ #split_tokens = [".", ":", ",",";","/","-","(", ")"]
147
+
148
+ TOKENIZER = Whitespace()
 
 
 
 
 
149
 
150
  def _info(self):
151
  # TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
 
280
  split_ids = np.array_split(clean_input_ids, n_chunks)
281
  split_labels = np.array_split(clean_labels, n_chunks)
282
  split_refs = np.array_split(clean_refs, n_chunks)
283
+ for chunk_ids, chunk_labels, chunk_refs in zip(clean_input_ids, clean_labels, clean_refs):
284
  # for chunk_id, index in enumerate(range(0, len(clean_input_ids), self.CHUNK_SIZE)):
285
  # split_ids = clean_input_ids[index:max(len(clean_input_ids), index + self.CHUNK_SIZE)]
286
  # split_bboxes = bboxes[index:index + self.CHUNK_SIZE]