Commits · bert-opt-layernorm · gaoqiong / MIGraphX

"script/cmake-cuda_docker.sh" did not exist on "bdbc0eaad175cd4054105cfb3fc812a8526e3b49"

23 Aug, 2022 12 commits
- Merge branch 'bert-opt-layernorm' of... · 54a4d80a
  Paul authored Aug 23, 2022
```
Merge branch 'bert-opt-layernorm' of github.com:ROCmSoftwarePlatform/AMDMIGraphX into bert-opt-layernorm
```
  54a4d80a
- Merge branch 'jit-layernorm-merge' into bert-opt-layernorm · b076d0f4
  Paul authored Aug 23, 2022
  
  b076d0f4
- Merge branch 'jit-layernorm-merge' of... · d705e483
  Paul authored Aug 23, 2022
```
Merge branch 'jit-layernorm-merge' of github.com:ROCmSoftwarePlatform/AMDMIGraphX into jit-layernorm-merge
```
  d705e483
- Format · ed34d352
  Paul authored Aug 23, 2022
  
  ed34d352
- Fix variance calculation · 3143befc
  Paul authored Aug 23, 2022
  
  3143befc
- Merge branch 'develop' into jit-layernorm-merge · 66d50268
  Paul Fultz II authored Aug 23, 2022
  
  66d50268
- Format · 389bc830
  Paul authored Aug 23, 2022
  
  389bc830
- Put into a detail namespace · 6c811ca5
  Paul authored Aug 23, 2022
  
  6c811ca5
- Handle implicit conversions · 22693358
  Paul authored Aug 23, 2022
  
  22693358
- Remove static_assert · 9f8c4774
  Paul authored Aug 23, 2022
  
  9f8c4774
- Merge branch 'improve-gpu-sched' into jit-layernorm-merge · 927580ef
  Paul authored Aug 23, 2022
  
  927580ef
- Dynamic ref NMS (#1288) · fa3c21fa
  Charlie Lin authored Aug 23, 2022
```
Has NMS op output a dynamic shape (ONNX spec behavior)
Allows for dynamic input shape to NMS op
```
  fa3c21fa
22 Aug, 2022 1 commit
- Vector dot · e5a25712
  Paul authored Aug 22, 2022
  
  e5a25712
21 Aug, 2022 5 commits
- Update is_supported (#1334) · 79e15ca9
  varunsh authored Aug 21, 2022
```
* Update is_supported
* Return object from is_supported
* Return by reference in interator
```
  79e15ca9
- Allow integral constant to be passed · c4add749
  Paul authored Aug 20, 2022
  
  c4add749
- Merge branch 'jit-layernorm-merge' into bert-opt-layernorm · 9cd72cdd
  Paul authored Aug 20, 2022
  
  9cd72cdd
- Format · 57a5c827
  Paul authored Aug 20, 2022
  
  57a5c827
- Use cast in non-constexpr call · 0f6c37e5
  Paul authored Aug 20, 2022
  
  0f6c37e5
20 Aug, 2022 12 commits
- Format · da786eb1
  Paul authored Aug 19, 2022
  
  da786eb1
- Use recip · 6c33a53f
  Paul authored Aug 19, 2022
  
  6c33a53f
- Format · 03c6967e
  Paul authored Aug 19, 2022
  
  03c6967e
- Use recipriocal · 07f37b4c
  Paul authored Aug 19, 2022
  
  07f37b4c
- Merge branch 'bert-opt-layernorm' of... · 319bac7c
  Paul authored Aug 19, 2022
```
Merge branch 'bert-opt-layernorm' of github.com:ROCmSoftwarePlatform/AMDMIGraphX into bert-opt-layernorm
```
  319bac7c
- Merge github.com:ROCmSoftwarePlatform/AMDMIGraphX into bert-opt-layernorm · ce19fbc6
  Paul authored Aug 19, 2022
  
  ce19fbc6
- Merge branch 'jit-layernorm-merge' into bert-opt-layernorm · 56a1645f
  Paul authored Aug 19, 2022
  
  56a1645f
- Merge · e7a1d5fa
  Paul authored Aug 19, 2022
  
  e7a1d5fa
- Format · d36f72c5
  Paul authored Aug 19, 2022
  
  d36f72c5
- Use vector types where possible · 8b5b6758
  Paul authored Aug 19, 2022
  
  8b5b6758
- Format · 73da1e4a
  Paul authored Aug 19, 2022
  
  73da1e4a
- Array foreach · dd900bf2
  Paul authored Aug 19, 2022
  
  dd900bf2
19 Aug, 2022 10 commits
- Format · 924a0ffa
  Paul authored Aug 19, 2022
  
  924a0ffa
- A · 73d53f86
  Paul authored Aug 19, 2022
  
  73d53f86
- Apply vec_reduce first · 8b8395be
  Paul authored Aug 19, 2022
  
  8b8395be
- Enable tidy for fpga backend (#1347) · b691abdd
  Paul Fultz II authored Aug 19, 2022
  
  b691abdd
- Reduce first · eb98f95b
  Paul authored Aug 19, 2022
  
  eb98f95b
- Adjust upper limit · 99c96638
  Paul authored Aug 19, 2022
  
  99c96638
- Merge branch 'jit-layernorm-merge' into bert-opt-layernorm · 06607821
  Paul authored Aug 19, 2022
  
  06607821
- Merge branch 'improve-gpu-sched' into bert-opt-layernorm · 6ef8bd98
  Paul authored Aug 19, 2022
  
  6ef8bd98
- Format · b5d703d8
  Paul authored Aug 19, 2022
  
  b5d703d8
- Use block size divisible by 32 and dont make global divisible by block size · 9ca5a6fc
  Paul authored Aug 19, 2022
  
  9ca5a6fc