vllm
[XPU] Initial support for GDN attention on Qwen3-next/Qwen3.5
#33657

Merged

Commits

[XPU] Support qwen3-next/qwen3.5 models

yma11 committed 126 days ago
fail lora path

yma11 committed 126 days ago
Need to align on block_size to 64 in TRITON_ATTN on XPU

yma11 committed 126 days ago
dispatch forward at initailization

yma11 committed 125 days ago
remove triton preferred_block_size

xuechendi committed 124 days ago
post process block_size for GDN on xpu

xuechendi committed 124 days ago
Merge pull request #2 from xuechendi/pr33657

yma11 committed 124 days ago
Merge branch 'main' into qwen3-next

jikunshang committed 124 days ago
update comments

yma11 committed 124 days ago