hf_text-generation-inference/server/exllamav2_kernels/setup.py

18 lines
465 B
Python

from setuptools import setup
from torch.utils.cpp_extension import BuildExtension, CUDAExtension
setup(
name="exllamav2_kernels",
ext_modules=[
CUDAExtension(
name="exllamav2_kernels",
sources=[
"exllamav2_kernels/ext.cpp",
"exllamav2_kernels/cuda/q_matrix.cu",
"exllamav2_kernels/cuda/q_gemm.cu",
],
)
],
cmdclass={"build_ext": BuildExtension},
)