Kernels:
Trusted publisher
|
Download README.md from kernels-community/flash-attn4: direct link, hf CLI and curl.
- Browser
- Download file 830 Bytes
-
https://huggingface.co/kernels/kernels-community/flash-attn4/resolve/v1/README.md
- Command line
-
hf download hf://kernels-community/flash-attn4@v1/README.md
-
curl -L -o README.md https://huggingface.co/kernels/kernels-community/flash-attn4/resolve/v1/README.md
830 Bytes
metadata
library_name: kernels
license: bsd-3-clause
This is the repository card of kernels-community/flash-attn4 that has been pushed on the Hub. It was built to be used with the kernels library. This card was automatically generated.
How to use
# make sure `kernels` is installed: `pip install -U kernels`
from kernels import get_kernel
kernel_module = get_kernel("kernels-community/flash-attn4", version=1)
flash_attn_func = kernel_module.flash_attn_func
flash_attn_func(...)
Available functions
flash_attn_funcflash_attn_varlen_func
Benchmarks
No benchmark available yet.
Source code
Source code of this kernel originally comes from https://github.com/Dao-AILab/flash-attention.git and it was repurposed for compatibility with kernels.