Tags: Distributed-Deep-Learning/byteps
Tags
packaging: ignore temp build directories (bytedance#319) - ignore temporary build directories in packaging. - check for existence of 3rdparty/ps-lite Signed-off-by: Yulu Jia <yulu.jia@bytedance.com>
Use true divide (bytedance#306) * torch: use tensor.true_divide() in the torch extension, for pytorch >= 1.5.0 use tensor.true_divide() (i.e. floating point divide) instead of tensor.div_() (i.e. integer divide) to average gradients. Signed-off-by: Yulu Jia <yulu.jia@bytedance.com>
bump version number to 0.2.5 (bytedance#305) bump version number to 0.2.5 Signed-off-by: Yulu Jia <yulu.jia@bytedance.com>
setup & docker: update framework and fix compile (bytedance#212) * setup: handle cxx flag * automatically add cuda lib for mxnet * fix softlink location * fix os path * fix ld_library_path * fix cuda default path * refactor dockerfilie * set proper nccl version * add torchvision * fix setup and use tf 1.15 for dockerfile * improve handling of libcuda.so.1
release: v0.2.0 (bytedance#206) * fix manifest.in * remove tools dir * __version__.py: 0.2.0 * update readme * fix docs * fix best-practice.md * readme: nccl path * readme: improve * readme: improve * add changelog.rst * fix changelog * readme: add pypi badge and news * improve readme and changelog