fix flash-attn comments (#354)

* Update README.md

* Update flash_attn_mma_share_kv.cu

* fix flash-attn comments

add out_f32x4_shared_bcf_merge_write_row2col(2d) (#339)

Bugfix: fix a compilation error (#336)

< < < should be <<<

Update FUNDING.yml

add triton merge_attn_states zhihu blog (#320)

Update README.md (#311)

feat: update pre-commit max-length=80 (#307)

* feat: update pre-commit length=120

* feat: update pre-commit max-length=80

Update README.md (#299)

feat: optimize merge_attn_states thread block dispatch (#279)

* [Kernel] opt cuda merge_attn_states kernel, part-1

* kernel: optimize merge_attn_states cuda kernel

[Docs] Add vLLM + DeepSeek-R1 671B deploy blog