[BugFix] Add embed_input_ids method to make QWenLMHeadModel a vllm model (#30674)

Signed-off-by: root <iwzbi@zju.edu.cn>
Co-authored-by: root <iwzbi@zju.edu.cn>
This commit is contained in:
duke
2025-12-15 10:38:29 +00:00
committed by GitHub
co-authored by root
parent 4429d934de
commit e4806d973a
+3
View File
@@ -281,6 +281,9 @@ class QWenBaseModel(nn.Module):
self.transformer.make_empty_intermediate_tensors
)
def embed_input_ids(self, input_ids: torch.Tensor) -> torch.Tensor:
return self.transformer.wte(input_ids)
def compute_logits(
self,
hidden_states: torch.Tensor,