Commits · aba8ead5bd5b7505fc95c4d138b96d2cd1100e04 · gaoqiong / MIGraphX

28 Nov, 2023 5 commits
- update eps · aba8ead5
  Umang Yadav authored Nov 28, 2023
  
  aba8ead5
- add pooling to unsupported ops · 3b40913c
  Umang Yadav authored Nov 28, 2023
  
  3b40913c
- fix conv_bn eps · a68536aa
  Umang Yadav authored Nov 28, 2023
  
  a68536aa
- add test for conv_bn with 1e-1f · 04efdb44
  Umang Yadav authored Nov 28, 2023
  
  04efdb44
- small example for fp8 fail case · 67a5ba86
  Umang Yadav authored Nov 28, 2023
  
  67a5ba86
27 Nov, 2023 2 commits
- add some MLIR fp8 tests for convolutions · 22aadbd6
  Umang Yadav authored Nov 27, 2023
  
  22aadbd6
- mlir fp8 · afe12765
  Umang Yadav authored Nov 26, 2023
  
  afe12765
26 Nov, 2023 11 commits
- fix cppcheck · c40a39c3
  Umang Yadav authored Nov 26, 2023
  
  c40a39c3
- formatting · 9a9e9648
  Umang Yadav authored Nov 26, 2023
  
  9a9e9648
- Move pass before optimize module · 83ce487a
  Umang Yadav authored Nov 26, 2023
  
  83ce487a
- Fix eliminate_fp8 pass · f014fb96
  Umang Yadav authored Nov 26, 2023
  
  f014fb96
- remove convert from lowering · 8734ffa3
  Umang Yadav authored Nov 26, 2023
  
  8734ffa3
- add eliminate_fp8 pass · ad9c25ea
  Umang Yadav authored Nov 26, 2023
  
  ad9c25ea
- add tests · 4604f2e1
  Umang Yadav authored Nov 26, 2023
  
  4604f2e1
- add tests and fix cmake · d473b802
  Umang Yadav authored Nov 26, 2023
  
  d473b802
- add ifdef to compile · 87548b5d
  Umang Yadav authored Nov 26, 2023
  
  87548b5d
- Works now · 037205c5
  Umang Yadav authored Nov 26, 2023
  
  037205c5
- compiles all right · 3aa465fd
  Umang Yadav authored Nov 26, 2023
  
  3aa465fd
21 Nov, 2023 2 commits
- add float8 for tensorview · a6c57726
  Umang Yadav authored Nov 21, 2023
  
  a6c57726
- use void for highest to match template candidates · 480288f8
  Umang Yadav authored Nov 21, 2023
  
  480288f8
20 Nov, 2023 10 commits
- Put numeric_max and numeeric lowest into float8 · f9542d5b
  Umang Yadav authored Nov 20, 2023
  
  f9542d5b
- Remove MIN max overloads · 836e201e
  Umang Yadav authored Nov 20, 2023
  
  836e201e
- Remove DPP · 2acd265b
  Umang Yadav authored Nov 20, 2023
  
  2acd265b
- Remove old cast · 6fbd9970
  Umang Yadav authored Nov 20, 2023
  
  6fbd9970
- add comment about c++20 extensions · 86605723
  Umang Yadav authored Nov 20, 2023
  
  86605723
- add test for rsqrt and remove old-styple-cast · 13ef4148
  Umang Yadav authored Nov 20, 2023
  
  13ef4148
- use __builtin_is_constant_evaluated · 6155c782
  Umang Yadav authored Nov 20, 2023
  
  6155c782
- tidy fixes · 7e3444ce
  Umang Yadav authored Nov 20, 2023
  
  7e3444ce
- add layernorm, remove constexpr for 1/r · f550f814
  Umang Yadav authored Nov 20, 2023
  
  f550f814
- roialign, softmax, pow, acosh, atanh,pad tests are enabled now · 13403ab2
  Umang Yadav authored Nov 20, 2023
  
  13403ab2
18 Nov, 2023 5 commits
- tidy fix · 1be95870
  Umang Yadav authored Nov 18, 2023
  
  1be95870
- add nolint · 249464c1
  Umang Yadav authored Nov 18, 2023
  
  249464c1
- fixes · 9f50051e
  Umang Yadav authored Nov 18, 2023
  
  9f50051e
- Fix tidy and other errors · e2b0c406
  Umang Yadav authored Nov 18, 2023
  
  e2b0c406
- Add math and reduce tests · 91cc9c7c
  Umang Yadav authored Nov 18, 2023
  
  91cc9c7c
17 Nov, 2023 5 commits
- add some more tests · bd0ae5fa
  Umang Yadav authored Nov 17, 2023
  
  bd0ae5fa
- Merge branch 'gpu_fp8' of https://github.com/ROCmSoftwarePlatform/AMDMIGraphX into gpu_fp8 · 76f03182
  Umang Yadav authored Nov 17, 2023
  
  76f03182
- revert unwanted changes · ef425d04
  Umang Yadav authored Nov 17, 2023
  
  ef425d04
- formatting · 60dd1f46
  Umang Yadav authored Nov 17, 2023
  
  60dd1f46
- Merge remote-tracking branch 'origin/develop' into gpu_fp8 · 3ae93ca9
  Umang Yadav authored Nov 17, 2023
  
  3ae93ca9