support transformers 4.34+ for llama (#9229)

intel-analytics · Oct 20, 2023 · 118249b · 118249b
1 parent 7e96d3e
commit 118249b
Showing 1 changed file with 2 additions and 0 deletions.
diff --git 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