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 5bf311637a fix(mge): fix infer output attrs fallible 3 years ago
..
amp refactor(imperative): move python code of elemwise/reduce/conv2d/bn to c++ 3 years ago
autodiff fix(mge/gm): fix missing dtype checking while attach tensors 3 years ago
core feat(mgb/dtype): add dtype qint1 3 years ago
data docs(imperative/dataloader): update preload description 3 years ago
distributed fix(mge/distributed): add polling to solve xmlrpc server io error 3 years ago
dtr refactor(imperative): remove command buffer 3 years ago
experimental refactor(dispatch): switch to new dispatch system 3 years ago
functional fix(mge): fix infer output attrs fallible 3 years ago
hub feat(imperative): mge hub pretrained support s3 3 years ago
jit feat(subgraph): subgraph builder supports jit and custom grad 3 years ago
module perf(imperative): add dim_expansion transform for conv/bn1d 3 years ago
optimizer perf(dispatch): speed up dispatch system 3 years ago
quantization docs(imperative): fix multiple target warning when cross-reference 3 years ago
random refactor(dispatch): switch to new dispatch system 3 years ago
tools Revert "fix(mge): replace _full_sync by sync" 3 years ago
traced_module perf(mge): move Const into C++ 3 years ago
utils fix(imperative): fix use collections error from python3.10 3 years ago
__init__.py feat(android/whl): support android whl 3 years ago
config.py fix(mge): control compute_mode by context 3 years ago
conftest.py ci(pytest): fixture doctest namespace in pytest config 3 years ago
device.py fix(imperative/cpu/multithread): fix multithread at imperative 3 years ago
logger.py fix(mge): fix some minor problems 3 years ago
serialization.py fix(imperative): fix the default pickle protocol version of save 3 years ago
tensor.py perf(dispatch): speed up dispatch system 3 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