ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

gguf-py : fix SpecialVocab parsing when post_processor is null (#14330)

Browse source
This commit is contained in:
Sigbjørn Skjæret 2025-06-22 19:46:17 +02:00 • committed by GitHub
parent 66aba7aca9
commit 238005c2dc
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 72 additions and 72 deletions
Download patch file Download diff file Expand all files Collapse all files

2
gguf-py/gguf/vocab.py
View file

@ -167,7 +167,7 @@ class SpecialVocab:
tokenizer_config['bos_token'] = special_bos = special_cls tokenizer_config['bos_token'] = special_bos = special_cls
if not special_eos and special_sep and tokenizer_config: if not special_eos and special_sep and tokenizer_config:
tokenizer_config['eos_token'] = special_eos = special_sep tokenizer_config['eos_token'] = special_eos = special_sep
post_processor = tokenizer.get('post_processor', {}) if post_processor := tokenizer.get('post_processor'):
for processor in post_processor.get('processors', [post_processor]): for processor in post_processor.get('processors', [post_processor]):
if processor.get('type') == 'RobertaProcessing': if processor.get('type') == 'RobertaProcessing':
self.add_special_token['bos'] = True self.add_special_token['bos'] = True