Kernels:
Trusted publisher
|
Download README.md from kernels-staging/finegrained-kernels: direct link, hf CLI and curl.
- Browser
- Download file 891 Bytes
-
https://huggingface.co/kernels/kernels-staging/finegrained-kernels/resolve/main/README.md
- Command line
-
hf download hf://kernels-staging/finegrained-kernels/README.md
-
curl -L -o README.md https://huggingface.co/kernels/kernels-staging/finegrained-kernels/resolve/main/README.md
891 Bytes
| library_name: kernels | |
| license: apache-2.0 | |
| This is the repository card of kernels-community/finegrained-kernels that has been pushed on the Hub. It was built to be used with the [`kernels` library](https://github.com/huggingface/kernels). This card was automatically generated. | |
| ## How to use | |
| ```python | |
| # make sure `kernels` is installed: `pip install -U kernels` | |
| from kernels import get_kernel | |
| kernel_module = get_kernel("kernels-community/finegrained-kernels", version=0) | |
| matmul_2d = kernel_module.matmul_2d | |
| matmul_2d(...) | |
| ``` | |
| ## Available functions | |
| - `matmul_2d` | |
| - `matmul_batched` | |
| - `matmul_grouped` | |
| - `moe_fused_batched` | |
| - `moe_fused_grouped` | |
| - `backward` | |
| - `get_supported_act_fns` | |
| - `get_supported_norms` | |
| - `rms_norm_rows` | |
| - `swizzle_mx_scales` | |
| - `unswizzle_mx_scales` | |
| - `mxfp8_act_quant` | |
| - `mxfp4_act_quant` | |
| - `nvfp4_act_quant` | |
| ## Benchmarks | |
| No benchmark available yet. | |