106 Commits (03b32e4eabc5128110ea9d673ba2ffc4bdd9cc2f)

Author SHA1 Message Date
  i-robot eaae9961a3
!30321 dynamic shape feed mode 4 years ago
  i-robot 4d1d38da81
!31734 fix mac undefined symbol error 4 years ago
  xulei 74e3552547 fix mac undefined symbol error 4 years ago
  i-robot 788365c2ea
!31719 Skip front cnode for device tensor store. 4 years ago
  gaoyong10 4e0dbffbc1 Skip front cnode for device tensor store. 4 years ago
  Henry Shi 2ade533b7f dynamic shape feed mode 4 years ago
  ttudu adb18b82ca dynamic shape CPU & GPU 4 years ago
  i-robot 5ca838b217
!31617 unified runtime support the link in the dynamic shape with heterogeneous 4 years ago
  limingqi107 9921444091 unified runtime support the link in the dynamic shape with heterogeneous 4 years ago
  i-robot 8997dd980f
!31406 Support online debugger for ascend mindRT kernel by kernel 4 years ago
  i-robot 048d089f9a
!31520 test ut 4 years ago
  gaoyong10 89c98f8b73 Add ut for dynamic shape and control flow. 4 years ago
  i-robot aaa9e262b8
!31527 fix the pclint of actor 4 years ago
  i-robot dc5f77f585
!31298 build libmindspore_backend.so 4 years ago
  limingqi107 1c39099072 fix the pclint of actor 4 years ago
  i-robot 83b51f8f5e
!31456 Support float immediate. 4 years ago
  gaoyong10 54fd41167a Support float immediate. 4 years ago
  zhoufeng b7bb53ff73 build libmindspore_backend.so 4 years ago
  i-robot ffaafc198f
!31351 fix the pclint of graph scheduler 4 years ago
  i-robot c3d15a079e
!31325 Training supports failover 4 years ago
  Parastoo Ashtari 018433e649 Support online debugger for ascend mindRT kernel by kernel 4 years ago
  lizhenyu 545707b80c Training supports failover 4 years ago
  i-robot db0ecf0c2d
!31340 Fix pclint. 4 years ago
  limingqi107 291a0b858c fix the pclint of graph scheduler 4 years ago
  hanhuifeng2020 089ce0386a [DynamicShape][Gpu]support dynamic shape for the DCN network 4 years ago
  gaoyong10 0c4ca0fefa Fix pclint. 4 years ago
  i-robot 6e4d2e0469
!31317 Fix pclint. 4 years ago
  i-robot 5716ad83e7
!29707 PyNative RunOp Async 4 years ago
  gaoyong10 1967ca2672 Fix pclint. 4 years ago
  i-robot a8d54beb5e
!31162 graph output address alloc free dynamically to support the dynamic shape sink mode 4 years ago
  caifubi af920290fb PyNative RunOp Async 4 years ago
  limingqi107 1ed1088cf7 graph output address alloc free dynamically to support the dynamic shape sink mode 4 years ago
  i-robot a33611ce27
!31192 Support sync and async message sending mode for TCP client 4 years ago
  Parallels 6f31f270e9 support sync and async message sending mode for tcp client 4 years ago
  ttudu 9e986fcedf fix input_node is data condaition 4 years ago
  i-robot 0376e934a5
!30943 PyNative RunOp without Actor 4 years ago
  caifubi d26d7b21ff PyNative RunOp without Actor 4 years ago
  i-robot a89faca8a0
!31115 unified runtime support the value node is output of subgraph in the heterogeneous scene 4 years ago
  gaoyong10 a94285147d Dynamic shape support control flow. 4 years ago
  limingqi107 75a7e00e2c unified runtime support the value node is output of subgraph in the heterogeneous scene 4 years ago
  i-robot 72777131c5
!30918 dynamic_shape fix bug 4 years ago
  i-robot 06feeaac62
!31047 Create device tensor with the device context of the kernel graph. 4 years ago
  ttudu 0069c82dab fix dynamic_shape bug 4 years ago
  gaoyong10 1ddd580949 Create device tensor with the device context of the kernel graph. 4 years ago
  gaoyong10 12f6decf2d Create kernel info for front node. 4 years ago
  i-robot 2ed51f3cc7
!30660 Add host_shape to CreateDeviceAddress API Input 4 years ago
  i-robot 911fb3995f
!30935 Collect device tensor in partial. 4 years ago
  i-robot 085d2dbe20
!30788 Add debug log and delete invalid code. 4 years ago
  i-robot 33495ae1e9
!30940 Use dynamic address for recv actor 4 years ago
  ZPaC 4d3ded56a9 Use dynamic address for recv actor 4 years ago