[Performance][Kernel] Fused_moe Performance Improvement #9384
comaniac
approved these changes
on 2024-10-18
add moe_sum and improve the thread number of moe align
34d4ea6a
linting
2f0fef9c
linting
7bd636f1
fix unit test
cf2c9900
charlifu
force pushed
to
cf2c9900
1 year ago
Merge branch 'vllm-project:main' into moe_opt
73fc6384
fix num_expert 64
4e5bd495
cast to in32
04af8129
charlifu
force pushed
to
04af8129
1 year ago
simon-mo
merged
59449095
into main 1 year ago
charlifu
deleted the moe_opt branch 1 year ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub