deepseek-ai / deepseek-ai/FlashMLA
why head_dim and vhead_dim only support 576,512
Open
- Dominant language
- C++
- Stars
- 12.9k
- Forks
- 1.2k
- Avg merge
- 4h 20m
- Merged PRs (30d)
- 2
Description
In the code, I noticed there are restrictions on head_dim and vhead_dim. What is the reason behind these constraints?
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.