| library_name: kernels | |
| license: mit | |
| This is the repository card of kernels-community/mlx-quantization-metal-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/mlx-quantization-metal-kernels", version=2) | |
| dequantize = kernel_module.dequantize | |
| dequantize(...) | |
| ``` | |
| ## Available functions | |
| - `dequantize` | |
| - `gather_qmm` | |
| - `gather_qqmm` | |
| - `qqmm` | |
| - `quantize` | |
| - `quantized_matmul` | |
| - `affine_gather_qmm_rhs_nax` | |
| - `affine_qmm_n` | |
| - `affine_qmm_n_nax` | |
| - `affine_qmm_t` | |
| - `affine_qmm_t_nax` | |
| - `affine_qmv` | |
| - `mxfp4_qmm_n` | |
| - `mxfp4_qmv` | |
| ## Benchmarks | |
| No benchmark available yet. | |