You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 
Megvii Engine Team 3bf73ff16f feat(dnn): add cuda preprocess fusion 5 years ago
..
adaptive_pooling feat(dnn/opr): add megdnn adaptive pooling opr 5 years ago
add_update MegEngine: Initial commit of MegEngine. 5 years ago
argmxx MegEngine: Initial commit of MegEngine. 5 years ago
argsort MegEngine: Initial commit of MegEngine. 5 years ago
batch_conv_bias refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
batch_normalization MegEngine: Initial commit of MegEngine. 5 years ago
batched_matrix_mul refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
checksum MegEngine: Initial commit of MegEngine. 5 years ago
concat MegEngine: Initial commit of MegEngine. 5 years ago
cond_take fix(dnn): generate mode files 5 years ago
conv_bias perf(dnn/cuda): change conv_bias heu, prefer dnn chanwise impl, dislike dnn batch gemm conv1x1 5 years ago
convolution refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
convolution3d refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
convolution_helper feat(windows/cuda/cmake): support cmake cuda build on windows 5 years ago
convpooling MegEngine: Initial commit of MegEngine. 5 years ago
cub MegEngine: Initial commit of MegEngine. 5 years ago
cumsum MegEngine: Initial commit of MegEngine. 5 years ago
cv MegEngine: Initial commit of MegEngine. 5 years ago
cvt_color MegEngine: Initial commit of MegEngine. 5 years ago
dct feat(dnn): add nchw_fp32 nchw44_qint8 cuda dct 5 years ago
deformable_conv refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
deformable_ps_roi_pooling feat(windows/cuda/cmake): support cmake cuda build on windows 5 years ago
dot MegEngine: Initial commit of MegEngine. 5 years ago
elemwise fix(dnn): generate mode files 5 years ago
elemwise_multi_type fix(dnn): generate mode files 5 years ago
eye MegEngine: Initial commit of MegEngine. 5 years ago
fake_quant feat(dnn/opr): add megdnn fake quant opr 5 years ago
flip MegEngine: Initial commit of MegEngine. 5 years ago
gaussian_blur MegEngine: Initial commit of MegEngine. 5 years ago
group_local refactor(dnn/cuda): invoke local opr in group local 5 years ago
images2neibs MegEngine: Initial commit of MegEngine. 5 years ago
indexing_multi_axis_vec fix(dnn): support bool for IndexingMultiAxisVec 5 years ago
indexing_one_hot MegEngine: Initial commit of MegEngine. 5 years ago
linspace MegEngine: Initial commit of MegEngine. 5 years ago
local refactor(dnn/cuda): remove function name weiming 5 years ago
local_share refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
lrn MegEngine: Initial commit of MegEngine. 5 years ago
mask_conv feat(dnn/naive): support weight preprocess interface in dnn 5 years ago
matrix_inverse MegEngine: Initial commit of MegEngine. 5 years ago
matrix_mul refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
max_tensor_diff MegEngine: Initial commit of MegEngine. 5 years ago
megcore MegEngine: Initial commit of MegEngine. 5 years ago
mesh_indexing MegEngine: Initial commit of MegEngine. 5 years ago
param_pack refactor(dnn/parampack): remove parampacksplit 5 years ago
pooling perf(dnn/cuda): enhance performance for pooling forward 5 years ago
powc MegEngine: Initial commit of MegEngine. 5 years ago
reduce fix(mgb/opr/reduce): fix grad function of reduce mean and add testcase 5 years ago
reduce_helper MegEngine: Initial commit of MegEngine. 5 years ago
relayout MegEngine: Initial commit of MegEngine. 5 years ago
relayout_format feat(dnn): add cuda preprocess fusion 5 years ago
remap feat(dnn/cuda): add remap backward 5 years ago
repeat MegEngine: Initial commit of MegEngine. 5 years ago
resize MegEngine: Initial commit of MegEngine. 5 years ago
rng MegEngine: Initial commit of MegEngine. 5 years ago
roi_align feat(mge): add bfloat16 support 5 years ago
roi_copy MegEngine: Initial commit of MegEngine. 5 years ago
roi_pooling feat(mge): add bfloat16 support 5 years ago
rotate MegEngine: Initial commit of MegEngine. 5 years ago
separable_conv MegEngine: Initial commit of MegEngine. 5 years ago
separable_filter MegEngine: Initial commit of MegEngine. 5 years ago
sleep MegEngine: Initial commit of MegEngine. 5 years ago
split MegEngine: Initial commit of MegEngine. 5 years ago
svd MegEngine: Initial commit of MegEngine. 5 years ago
tensor_remap feat(megbrain/dnn): add indexing remap int32 for naive and cuda 5 years ago
tile MegEngine: Initial commit of MegEngine. 5 years ago
topk perf(mgb/cuda): do not call cudaGetDeviceProperties to avoid io traffic 5 years ago
transpose MegEngine: Initial commit of MegEngine. 5 years ago
type_cvt fix(dnn): fix bool cvt 5 years ago
warp_affine MegEngine: Initial commit of MegEngine. 5 years ago
warp_perspective feat(dnn/cuda): add warp perspective backward mat idx 5 years ago
winograd_filter_preprocess MegEngine: Initial commit of MegEngine. 5 years ago
cuda_shfl_compat.cuh MegEngine: Initial commit of MegEngine. 5 years ago
cudnn_with_check.h MegEngine: Initial commit of MegEngine. 5 years ago
cudnn_wrapper.cpp refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
cudnn_wrapper.h refactor(dnn): refactor algo interface, use algoinfo instead of global algorithm 5 years ago
elemwise_helper.cpp feat(dnn): add bool dtype 5 years ago
elemwise_helper.cuh feat(dnn/opr): add megdnn fake quant opr 5 years ago
error_info.cuh MegEngine: Initial commit of MegEngine. 5 years ago
fp16_help.cuh MegEngine: Initial commit of MegEngine. 5 years ago
handle.cpp refactor(mgb/cuda): use single implementation of get_device_prop from utils 5 years ago
handle.h refactor(mgb/cuda): use single implementation of get_device_prop from utils 5 years ago
handle_create.cpp feat(dnn/opr): add megdnn fake quant opr 5 years ago
int_fastdiv.cpp MegEngine: Initial commit of MegEngine. 5 years ago
int_fastdiv.cuh MegEngine: Initial commit of MegEngine. 5 years ago
query_blocksize.cpp MegEngine: Initial commit of MegEngine. 5 years ago
query_blocksize.cuh MegEngine: Initial commit of MegEngine. 5 years ago
query_blocksize_impl.cu MegEngine: Initial commit of MegEngine. 5 years ago
reduce_helper.cuh MegEngine: Initial commit of MegEngine. 5 years ago
reduce_helper.cuinl MegEngine: Initial commit of MegEngine. 5 years ago
utils.cpp refactor(mgb/cuda): use single implementation of get_device_prop from utils 5 years ago
utils.cuh feat(dnn/cuda): integrate cutlass nchw32 tensorcore convolution 5 years ago
utils.h refactor(mgb/cuda): use single implementation of get_device_prop from utils 5 years ago

MegEngine 安装包中集成了使用 GPU 运行代码所需的 CUDA 环境,不用区分 CPU 和 GPU 版。 如果想要运行 GPU 程序,请确保机器本身配有 GPU 硬件设备并安装好驱动。 如果你想体验在云端 GPU 算力平台进行深度学习开发的感觉,欢迎访问 MegStudio 平台

C++ Cuda Python C SVG other