Zhiyuan Li

icon indicating a website link https://blog.uniartisan.com

icon company University of Melbourne

Results 63 comments of Zhiyuan Li

[RFC] Supports dispatch from multiple backend kernels

Triton, Tilelang, CuteTile, CuteDSL, and CUDA C could be worth considering.

[Linear Attention] Update fused_recurrent.py for inference with nomalization=true

Thanks for contributing, can you @yiyousong add tests to your contribution? This will improve the robustness of the code @yzhangcs could you please give some comments? tests: https://github.com/fla-org/flash-linear-attention/blob/main/tests/ops/test_linear_attn.py layers: https://github.com/fla-org/flash-linear-attention/blob/main/fla/layers/linear_attn.py

[Linear Attention] Update fused_recurrent.py for inference with nomalization=true

> > Thanks for contributing, can you @yiyousong add tests to your contribution? This will improve the robustness of the code > > @yzhangcs could you please give some comments?...