youngkingdom/vllm
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[Kernel] Marlin Expansion: Support AutoGPTQ Models with Marlin (#3922)

Browse Source 
Co-authored-by: alexm <alexm@neuralmagic.com>
Co-authored-by: mgoin <michael@neuralmagic.com>
This commit is contained in:
Robert Shaw
2024-04-29 12:35:34 -04:00
committed by GitHub
parent df29793dc7
commit 73c8d677e5
14 changed files with 2627 additions and 105 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
csrc/pybind.cpp
View File

@ -67,6 +67,8 @@ PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) {
ops.def("aqlm_dequant", &aqlm_dequant, "Decompression method for AQLM");
ops.def("awq_gemm", &awq_gemm, "Quantized GEMM for AWQ");
ops.def("marlin_gemm", &marlin_gemm, "Marlin Optimized Quantized GEMM for GPTQ");
ops.def("gptq_marlin_gemm", &gptq_marlin_gemm, "gptq_marlin Optimized Quantized GEMM for GPTQ");
ops.def("gptq_marlin_repack", &gptq_marlin_repack, "gptq_marlin repack from GPTQ");
ops.def("awq_dequantize", &awq_dequantize, "Dequantization for AWQ");
#endif