xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-06-05 11:29:09 +08:00
Code Issues Packages Projects Releases Wiki Activity

[bugfix] remove unused parameters to reduce unnecessary vram usage (#26789)

Browse Source 
Signed-off-by: Reinforce-II <fate@eastal.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
This commit is contained in:
Reinforce-II 2025-10-22 23:16:09 +08:00 committed by GitHub
parent 1c160841ea
commit 980de31ca0
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 2 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors_moe.py
View File

@ -307,10 +307,12 @@ class CompressedTensorsW4A4MoeMethod(CompressedTensorsMoEMethod):
layer.w13_weight = torch.nn.Parameter( layer.w13_weight = torch.nn.Parameter(
layer.w13_weight_packed.data, requires_grad=False layer.w13_weight_packed.data, requires_grad=False
) )
delattr(layer, "w13_weight_packed")
layer.w2_weight = torch.nn.Parameter( layer.w2_weight = torch.nn.Parameter(
layer.w2_weight_packed.data, requires_grad=False layer.w2_weight_packed.data, requires_grad=False
) )
delattr(layer, "w2_weight_packed")
# reorder GEMM1 weights and block scales for FlashInfer CUTLASS kernel. # reorder GEMM1 weights and block scales for FlashInfer CUTLASS kernel.
if self.allow_flashinfer: if self.allow_flashinfer: