xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-05-14 17:47:57 +08:00
Code Issues Packages Projects Releases Wiki Activity

[BugFix] Fix DeepGEMM warmup, no m.weight_scale_inv (#25206)

Browse Source 
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
This commit is contained in:
Lucas Wilkinson 2025-09-18 17:26:28 -04:00 committed by GitHub
parent a53ad626d6
commit 9fac6aa30b
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/warmup/deep_gemm_warmup.py
View File

@ -36,7 +36,7 @@ def _extract_data_from_linear_base_module(
assert m.quant_method.quant_config is not None assert m.quant_method.quant_config is not None
w = m.weight w = m.weight
ws = m.weight_scale_inv ws = m.weight_scale
quant_block_size = m.quant_method.quant_config.weight_block_size quant_block_size = m.quant_method.quant_config.weight_block_size
assert isinstance(w, torch.Tensor) assert isinstance(w, torch.Tensor)