| .. |
|
all_reduce
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
attention
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
backup
|
f8dfac6372
chore: attention refactor and upstream sync apr01 (#365)
|
11 ヶ月 前 |
|
cpu
|
8c32e49029
feat: add avx2 cpu support
|
7 ヶ月 前 |
|
hadamard
|
5d288aa76c
feat: add fast hadamard transformation kernels (#232)
|
1 年間 前 |
|
moe
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
punica
|
da6765c084
feat: lora support for commandr models
|
7 ヶ月 前 |
|
quantization
|
cd9ed8623b
fix: cuda version check for fp8 support in the cutlass kernels
|
7 ヶ月 前 |
|
activation_kernels.cu
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
cache.h
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
cache_kernels.cu
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
cuda_compat.h
|
00acf371f9
rocm: fused topk softmax
|
7 ヶ月 前 |
|
cuda_utils.h
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
cuda_utils_kernels.cu
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
dispatch_utils.h
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
layernorm_kernels.cu
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
ops.h
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
pos_encoding_kernels.cu
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
reduction.cuh
|
aba03b4756
feat: dynamic per-token activation quantization
|
7 ヶ月 前 |
|
registration.h
|
156f577f79
feat: switch from `PYBIND11_MODULE` to `TORCH_LIBRARY` (#569)
|
7 ヶ月 前 |
|
torch_bindings.cpp
|
cd9ed8623b
fix: cuda version check for fp8 support in the cutlass kernels
|
7 ヶ月 前 |
AlpinDale
cd9ed8623b
fix: cuda version check for fp8 support in the cutlass kernels