This commit is contained in:
turboderp
2024-07-08 03:51:14 +02:00
parent 8bbce3455c
commit e97ad9cb27

View File

@@ -61,7 +61,8 @@ def _get_lmfe_tokenizer_data(tokenizer: ExLlamaV2Tokenizer):
def clear_grammar_func_cache():
"""Flush tokenizer_data cache to avoid holding references to tokenizers after unloading a model"""
"""Flush tokenizer_data cache to avoid holding references to
tokenizers after unloading a model"""
_get_lmfe_tokenizer_data.clear_cache()