ctheodoris commited on
Commit
0f1a805
1 Parent(s): 43bb5ab

account for more than 1 gene being overexpressed

Browse files
Files changed (1) hide show
  1. geneformer/perturber_utils.py +1 -1
geneformer/perturber_utils.py CHANGED
@@ -259,7 +259,7 @@ def overexpress_tokens(example, max_len, special_token):
259
  # truncate to max input size, must also truncate original emb to be comparable
260
  if len(example["input_ids"]) > max_len:
261
  if special_token:
262
- del example["input_ids"][max_len-1]
263
  else:
264
  example["input_ids"] = example["input_ids"][0:max_len]
265
  example["length"] = len(example["input_ids"])
 
259
  # truncate to max input size, must also truncate original emb to be comparable
260
  if len(example["input_ids"]) > max_len:
261
  if special_token:
262
+ example["input_ids"] = example["input_ids"][0:max_len-1]+[example["input_ids"][-1]]
263
  else:
264
  example["input_ids"] = example["input_ids"][0:max_len]
265
  example["length"] = len(example["input_ids"])