tommasobaldi commited on
Commit
530ce31
·
1 Parent(s): 9c2c31c

working on text splitting

Browse files
Files changed (1) hide show
  1. app.py +1 -1
app.py CHANGED
@@ -104,7 +104,7 @@ def main() -> None:
104
  result = []
105
  for sentence in sentences:
106
  tokens = word_tokenize(sentence)
107
- if token_count + len(tokens) < 1024:
108
  token_count += len(tokens)
109
  text_block += " ".join(sentence)
110
  else:
 
104
  result = []
105
  for sentence in sentences:
106
  tokens = word_tokenize(sentence)
107
+ if token_count + len(tokens) < 500:
108
  token_count += len(tokens)
109
  text_block += " ".join(sentence)
110
  else: