Skip to content

Commit

Permalink
Merge pull request vllm-project#7 from ri938/remove_fies
Browse files Browse the repository at this point in the history
remove not needed files
  • Loading branch information
ri938 authored Aug 24, 2023
2 parents 73db30f + eed1888 commit aaea899
Show file tree
Hide file tree
Showing 7 changed files with 0 additions and 301 deletions.
113 changes: 0 additions & 113 deletions vllm/awq_quantization/kernels/csrc/layernorm/layernorm.cu

This file was deleted.

3 changes: 0 additions & 3 deletions vllm/awq_quantization/kernels/csrc/layernorm/layernorm.h

This file was deleted.

82 changes: 0 additions & 82 deletions vllm/awq_quantization/kernels/csrc/layernorm/reduction.cuh

This file was deleted.

This file was deleted.

This file was deleted.

4 changes: 0 additions & 4 deletions vllm/awq_quantization/kernels/csrc/pybind.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -2,13 +2,9 @@

#include <pybind11/pybind11.h>
#include <torch/extension.h>
#include "layernorm/layernorm.h"
#include "quantization/gemm_cuda.h"
#include "position_embedding/pos_encoding.h"

PYBIND11_MODULE(TORCH_EXTENSION_NAME, m)
{
m.def("layernorm_forward_cuda", &layernorm_forward_cuda, "FasterTransformer layernorm kernel");
m.def("gemm_forward_cuda", &gemm_forward_cuda, "Quantized GEMM kernel.");
m.def("rotary_embedding_neox", &rotary_embedding_neox, "Apply GPT-NeoX style rotary embedding to query and key");
}
2 changes: 0 additions & 2 deletions vllm/awq_quantization/kernels/setup.py
Original file line number Diff line number Diff line change
Expand Up @@ -17,8 +17,6 @@
sources=[
"csrc/pybind.cpp",
"csrc/quantization/gemm_cuda_gen.cu",
"csrc/layernorm/layernorm.cu",
"csrc/position_embedding/pos_encoding_kernels.cu"
],
extra_compile_args=extra_compile_args,
),
Expand Down

0 comments on commit aaea899

Please sign in to comment.