ayo/ipex-llm
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 1

fix in transformers 4.36 (#10150)

Browse source
This commit is contained in:
Jiao Wang 2024-02-28 18:43:01 -08:00 committed by GitHub
parent 43dac97e03
commit 6fb65bb9d2
1 changed files with 1 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
python/llm/portable-zip/chat.py
View file

@ -288,7 +288,7 @@ if __name__ == "__main__":
if model.config.architectures is not None and model.config.architectures[0] == "QWenLMHeadModel": if model.config.architectures is not None and model.config.architectures[0] == "QWenLMHeadModel":
stop_words = get_stop_words_ids("Qwen", tokenizer=tokenizer) stop_words = get_stop_words_ids("Qwen", tokenizer=tokenizer)
kv_cache = StartRecentKVCache(start_size=start_size) kv_cache = StartRecentKVCache(start_size=start_size, k_seq_dim=1, v_seq_dim=1)
qwen_stream_chat(model=model, tokenizer=tokenizer,kv_cache=kv_cache, stop_words=stop_words) qwen_stream_chat(model=model, tokenizer=tokenizer,kv_cache=kv_cache, stop_words=stop_words)
elif model.config.architectures is not None and model.config.architectures[0] == "ChatGLMModel": elif model.config.architectures is not None and model.config.architectures[0] == "ChatGLMModel":
chatglm3_stream_chat(model=model, tokenizer=tokenizer) chatglm3_stream_chat(model=model, tokenizer=tokenizer)