fix xverse

Signed-off-by: Roger Wang <[email protected]>
vllm-project · Nov 16, 2024 · 719e34c · 719e34c
1 parent e92d165
commit 719e34c
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/vllm/model_executor/models/xverse.py b/vllm/model_executor/models/xverse.py
@@ -352,9 +352,11 @@ def forward(
         kv_caches: List[torch.Tensor],
         attn_metadata: AttentionMetadata,
         intermediate_tensors: Optional[IntermediateTensors] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
     ) -> Union[torch.Tensor, IntermediateTensors]:
         hidden_states = self.model(input_ids, positions, kv_caches,
-                                   attn_metadata, intermediate_tensors)
+                                   attn_metadata, intermediate_tensors,
+                                   inputs_embeds)
         return hidden_states
 
     def compute_logits(