1. 11 4月, 2018 1 次提交
  2. 19 3月, 2018 3 次提交
  3. 18 3月, 2018 1 次提交
  4. 17 3月, 2018 1 次提交
  5. 12 2月, 2018 1 次提交
  6. 10 2月, 2018 2 次提交
  7. 25 12月, 2017 1 次提交
  8. 22 12月, 2017 1 次提交
  9. 12 12月, 2017 1 次提交
    • Q
      Refine device context (#6433) · 61ec0b95
      QI JUN 提交于
      There are mainly following fixes:
      
      - take `DeviceContext` as the template parameter of math functors and OpKernel instead of `Place`
      - remove `eigen_device` interface in base class  `DeviceContext`
      - remove `GetEigenDevice` interface in `ExecutionContext` and base class `DeviceContext`
      - remove unused `platform::EigenDeviceConverter`
      - rename `REGISTER_OP_GPU_KERNEL` to `REGISTER_OP_CUDA_KERNEL`
      - rename `USE_GPU_ONLY_OP` to `USE_CUDA_ONLY_OP`
      61ec0b95
  10. 28 11月, 2017 1 次提交
  11. 11 11月, 2017 1 次提交
  12. 28 10月, 2017 2 次提交
  13. 26 10月, 2017 1 次提交
    • Q
      Cudnn batch norm op (#5067) · 56b723c4
      Qiao Longfei 提交于
      * init cudnn batch norm op
      
      * rename batch_norm_cudnn_op.cc batch_norm_op.cu
      
      * correct name style
      
      * add ExtractNCWHD, simplify code
      
      * fix ExtractNCWHD
      
      * use CUDNN_ENFORCE instead of PADDLE_ENFORCE
      56b723c4