A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 03:30 UTC
[Model] Implement missing `get_language_model` for Keye-VL (#20631)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Cyrus Leung committed
Jul 8, 2025 at 22:47 UTC
dd382e0fe377b839189cd27db93f11a6cfe35250
1 file changed
+3
vllm/model_executor/models/keye.py
+3
@@ -1592,6 +1592,9 @@ class KeyeForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsLoRA,
1592
1593
return modalities
1594
1595
+ def get_language_model(self) -> torch.nn.Module:
1596
+ return self.language_model
1597
+
1598
def get_multimodal_embeddings(
1599
self, **kwargs: object) -> Optional[MultiModalEmbeddings]:
1600