220 Commits (bc738d4ec3e73f4e25a3631f71b9833fe6830e47)

Author SHA1 Message Date
  mindspore-ci-bot 8a8de7e062 !3171 gpu fix the graph of 'nop node + depend + node' 5 years ago
  laiyongqiang b570dec7ab add right align border for communication op's single output 5 years ago
  limingqi107 a596dd6e43 gpu fix the graph of 'nop node + depend + node' 5 years ago
  yujianfeng 188d74f15e Remove transdata and cast for internal outputs 5 years ago
  mindspore-ci-bot 7233d650f0 !3063 Enable to train in parameter server mode 5 years ago
  ZPaC 52022c8013 Enable to train in parameter server mode 5 years ago
  mindspore-ci-bot ae50c37c38 !3092 GPU add fuison: replace momentum cast 5 years ago
  VectorSL 140174182d gpu add fusion: replace momentum cast 5 years ago
  mindspore-ci-bot b64fca6e05 !3091 GPU add fusion: replace batchnormgrad cast 5 years ago
  mindspore-ci-bot 0c7811522f !3090 GPU add fusion 5 years ago
  changzherui f4cb445ea8 syn code for 0715 5 years ago
  mindspore-ci-bot a581766be4 !3069 Enable optimizer parallel with broadcast 5 years ago
  mindspore-ci-bot 25ee322ba3 !2966 reuse communication op output's memory 5 years ago
  laiyongqiang 68c78ab6bb reuse communication op output's memory 5 years ago
  Ziyan 39f08eb7dd enable optimizer parallel 5 years ago
  VectorSL 4cf7faeae6 gpu add fusion: replace batchnorm grad cast 5 years ago
  VectorSL 072b09b3fd gpu add fusion: 1 replace bn cast 2 replace addn by tensoradd 5 years ago
  lichen_101010 6bb2182134 Add partial memory reuse support to debugger 5 years ago
  He Wei ee1510da41 Eliminate circular dependency between 'ir' and 'device/kernel' 5 years ago
  liubuyu 43c79eb853 mindspore path adjust 5 years ago