Instructions to use replicate/deep-gemm with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Kernels
How to use replicate/deep-gemm with Kernels:
# !pip install kernels from kernels import get_kernel # a version (or an explicit revision) is required; see the "Files and versions" tab for the available ones kernel = get_kernel("replicate/deep-gemm", version=1) - Notebooks
- Google Colab
- Kaggle
Download torch-ext/deep_gemm/__init__.py from replicate/deep-gemm: direct link, hf CLI and curl.
- Browser
- Download file 351 Bytes
-
https://huggingface.co/replicate/deep-gemm/resolve/main/torch-ext/deep_gemm/__init__.py
- Command line
-
hf download hf://replicate/deep-gemm/torch-ext/deep_gemm/__init__.py
-
curl -L -o __init__.py https://huggingface.co/replicate/deep-gemm/resolve/main/torch-ext/deep_gemm/__init__.py
351 Bytes
| import torch | |
| from . import jit | |
| from .jit_kernels import ( | |
| gemm_fp8_fp8_bf16_nt, | |
| m_grouped_gemm_fp8_fp8_bf16_nt_contiguous, | |
| m_grouped_gemm_fp8_fp8_bf16_nt_masked, | |
| ceil_div, | |
| set_num_sms, get_num_sms, | |
| get_col_major_tma_aligned_tensor, | |
| get_m_alignment_for_contiguous_layout | |
| ) | |
| from .utils import bench, bench_kineto, calc_diff | |