1. 22 7月, 2020 1 次提交
    • H
      [Core] Add the graph optimization of subblocks for transformer model (#3947) · 7af1a258
      hong19860320 提交于
      * [Core][ARM] Fix beam_search, eltwise_mul supports broadcast and int64_t data type, add print op and kernel, add exeception
      test=develop
      
      * Fix the dims of parent idx of the arm kernel of beam_search op
      
      * elementwise_mul supports int64_t data type with broadcasting
      
      * Add print op and kernel for debugging
      
      * Support throwing the exception when the internal error occurs
      
      * Refine while and conditional_block op kernel
      
      * Support the graph optimization on subblocks
      
      * Pass program_desc and block_idx into the kernel of the control flow ops(while/conditional_block/subgraph), and create the RuntimeProgram online, it make it possiable to call the control flow ops recursively
      
      *Add unit test for masked transformer model
      7af1a258
  2. 10 7月, 2020 1 次提交
  3. 07 7月, 2020 1 次提交
  4. 01 7月, 2020 1 次提交
  5. 29 6月, 2020 1 次提交
  6. 18 5月, 2020 1 次提交
  7. 12 5月, 2020 1 次提交
  8. 20 4月, 2020 1 次提交
  9. 08 4月, 2020 1 次提交
  10. 03 4月, 2020 1 次提交
  11. 31 3月, 2020 1 次提交
  12. 25 3月, 2020 1 次提交
  13. 11 10月, 2019 1 次提交
    • Z
      CUDA: can run yolov3 int8 (#2172) · 7931104f
      Zhaolong Xing 提交于
      * add conv int8 support(in condition which the input or output channel not be the times of 4)
      add add_kernel for cuda.
      
      * can run yolov3 fp32
      test=develop
      
      * 1. fix bug with yolov3 run
      test=develop
      
      * can run yolov3 int8 test=develop
      7931104f
  14. 27 9月, 2019 1 次提交
  15. 16 8月, 2019 1 次提交