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, 04:45 UTC
[BUGFIX][DEEPSEEK][MODEL_LOAD] fix w13, w2 weight not initialized assert (#20202)
Signed-off-by: Chendi Xue <chendi.xue@intel.com>
Chendi.Xue committed
Jun 29, 2025 at 21:46 UTC
5a52f389dde444b9e122a7fa903393fd64857b86
1 file changed
+1
vllm/model_executor/models/deepseek_v2.py
+1
@@ -889,6 +889,7 @@ class DeepseekV2ForCausalLM(nn.Module, SupportsPP, MixtureOfExperts):
889
expert_id=expert_id,
890
return_success=True)
891
if success:
892
+ name = name_mapped
893
break
894
else:
895
if is_expert_weight: