Commits · 3119fa01d4c9f51434f20d1d4c54b409ab442a05 · gaoqiong / MIGraphX

10 Jun, 2019 2 commits
- Merge branch 'int8_miopen_call' into op_capture · 3119fa01
  Shucai Xiao authored Jun 09, 2019
  
  3119fa01
- fixed a bug with block size · e8858300
  Shucai Xiao authored Jun 09, 2019
  
  e8858300
07 Jun, 2019 1 commit
- merge changes from develop branch. · 40a5f496
  Shucai Xiao authored Jun 07, 2019
  
  40a5f496
06 Jun, 2019 12 commits
- Merge branch 'develop' of https://github.com/ROCmSoftwarePlatform/AMDMIGraphX into op_capture · 0379dd66
  Shucai Xiao authored Jun 06, 2019
  
  0379dd66
- Merge pull request #267 from ROCmSoftwarePlatform/conv_same_padding · 5a33a3f9
  mvermeulen authored Jun 06, 2019
```
padding calculation for attribute
```
  5a33a3f9
- clang format · fb5df2fc
  Shucai Xiao authored Jun 06, 2019
  
  fb5df2fc
- add an overload for function capture_arguments · 6177d2b3
  Shucai Xiao authored Jun 06, 2019
  
  6177d2b3
- formatting · 4fd8c544
  Khalique authored Jun 06, 2019
  
  4fd8c544
- manual merge · 15a79fd9
  Khalique authored Jun 06, 2019
  
  15a79fd9
- Merge branch 'int8_quantize' into op_capture · 85e4d7ac
  Shucai Xiao authored Jun 06, 2019
  
  85e4d7ac
- merge changes from develop branch. · f8613dd1
  Shucai Xiao authored Jun 06, 2019
  
  f8613dd1
- add a python interface for int8 quantization. · 09740556
  Shucai Xiao authored Jun 06, 2019
  
  09740556
- Merge pull request #264 from ROCmSoftwarePlatform/hip_softmax · 0d52d99f
  mvermeulen authored Jun 06, 2019
```
Add hip softmax
```
  0d52d99f
- clang format · 86bfe27c
  Shucai Xiao authored Jun 05, 2019
  
  86bfe27c
- fixed a race condition problem in quant_gemm. · 9174b783
  Shucai Xiao authored Jun 05, 2019
  
  9174b783
05 Jun, 2019 16 commits
- Merge branch 'int8_quantize' into op_capture · 277ae59c
  Shucai Xiao authored Jun 05, 2019
  
  277ae59c
- clang format · 4cf9bd01
  Shucai Xiao authored Jun 05, 2019
  
  4cf9bd01
- fix cpu and gpu execution mismatch. · 7f9e01d0
  Shucai Xiao authored Jun 05, 2019
  
  7f9e01d0
- clang format · 43d39b4e
  Shucai Xiao authored Jun 05, 2019
  
  43d39b4e
- fixe mismatch between cpu and gpu execution. · ffa6a45a
  Shucai Xiao authored Jun 05, 2019
  
  ffa6a45a
- reinsert miopen_softmax · e1758782
  Khalique authored Jun 05, 2019
  
  e1758782
- add quantize_int8 python interace. · e6158f10
  Shucai Xiao authored Jun 05, 2019
  
  e6158f10
- add the overload quantize_int8 function api to the quantization head file. · 11a0f1a9
  Shucai Xiao authored Jun 05, 2019
  
  11a0f1a9
- add python interface for quantize int8. · a8fadc6f
  Shucai Xiao authored Jun 05, 2019
  
  a8fadc6f
- add python interface for the capture_argument function. · c63d2dd3
  Shucai Xiao authored Jun 05, 2019
  
  c63d2dd3
- clang format · 58ff9509
  Shucai Xiao authored Jun 05, 2019
  
  58ff9509
- merge changes from int8_quantize · 0b2fb46a
  Shucai Xiao authored Jun 05, 2019
  
  0b2fb46a
- code cleanup · 2fcc8c09
  Shucai Xiao authored Jun 05, 2019
  
  2fcc8c09
- Merge branch 'develop' of https://github.com/ROCmSoftwarePlatform/AMDMIGraphX into op_capture · 0d5aa0f1
  Shucai Xiao authored Jun 05, 2019
  
  0d5aa0f1
- clang format · 4d1f18a8
  Shucai Xiao authored Jun 05, 2019
  
  4d1f18a8
- simply the quantize int8 implementation. · c694a4d6
  Shucai Xiao authored Jun 05, 2019
  
  c694a4d6
04 Jun, 2019 9 commits
- add std modifier · 33d6d426
  Khalique authored Jun 04, 2019
  
  33d6d426
- clang format · 54cbce68
  Shucai Xiao authored Jun 04, 2019
  
  54cbce68
- change quantize_int8 following changes of quant_convolution. · 675d3b5b
  Shucai Xiao authored Jun 04, 2019
  
  675d3b5b
- move header file · 88dac3b8
  Khalique authored Jun 04, 2019
  
  88dac3b8
- Merge branch 'int8_miopen_call' into int8_quantize · 6e01c7cb
  Shucai Xiao authored Jun 04, 2019
  
  6e01c7cb
- Merge branch 'develop' of https://github.com/ROCmSoftwarePlatform/AMDMIGraphX into int8_quantize · 13dfa914
  Shucai Xiao authored Jun 04, 2019
  
  13dfa914
- clang format · ee157fc2
  Shucai Xiao authored Jun 04, 2019
  
  ee157fc2
- fake up to make the int8_convolution outputs the int32_t data type. With... · 870f565e
  Shucai Xiao authored Jun 04, 2019
```
fake up to make the int8_convolution outputs the int32_t data type. With MIOpen to support int32_t output shape, this tmp change will be removed.
```
  870f565e
- Merge branch 'develop' of https://github.com/ROCmSoftwarePlatform/AMDMIGraphX into int8_miopen_call · dbd34685
  Shucai Xiao authored Jun 03, 2019
  
  dbd34685