From 118249b011e96632d76a202cfd835c7e0466c269 Mon Sep 17 00:00:00 2001 From: Yang Wang Date: Fri, 20 Oct 2023 13:36:30 +0800 Subject: [PATCH] support transformers 4.34+ for llama (#9229) --- python/llm/src/bigdl/llm/transformers/models/llama.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/python/llm/src/bigdl/llm/transformers/models/llama.py b/python/llm/src/bigdl/llm/transformers/models/llama.py index 7953670a..0dd39ae6 100644 --- a/python/llm/src/bigdl/llm/transformers/models/llama.py +++ b/python/llm/src/bigdl/llm/transformers/models/llama.py @@ -79,6 +79,8 @@ def llama_attention_forward_4_31( past_key_value: Optional[Tuple[torch.Tensor]] = None, output_attentions: bool = False, use_cache: bool = False, + padding_mask: Optional[torch.LongTensor] = None, + **kwargs, ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]: bsz, q_len, _ = hidden_states.size() device = hidden_states.device