biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Feature] Add SM103 (Blackwell Ultra) Support to vLLM (#30484)

Browse Source 
Signed-off-by: LopezCastroRoberto <robertol.c510@gmail.com>
Signed-off-by: Roberto L. Castro <38211239+LopezCastroRoberto@users.noreply.github.com>
Co-authored-by: youkaichao <youkaichao@gmail.com>
This commit is contained in:
Roberto L. Castro
2025-12-13 04:34:23 +01:00
committed by GitHub
parent 57e9bf1864
commit 4fa7ce46f3
21 changed files with 53 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/models/config.py
View File

@@ -363,7 +363,7 @@ class HybridAttentionMambaModelConfig(VerifyAndUpdateConfig):
else:
kernel_block_alignment_size = 16
if (
current_platform.is_device_capability(100)
current_platform.is_device_capability_family(100)
and model_config.get_head_size() == 256
and (
attention_config.backend is None