Popular repositories
-
Forked from HowardHinnant/date
A date and time library based on the C++11/14/17 <chrono> header
C++ 1
-
-
Forked from Ciantic/VirtualDesktopAccessor
DLL for accessing Windows 10 Virtual Desktop features from e.g. AutoHotkey
C++
-
-
688 contributions in the last year
Less
More
Activity overview
Contributed to
microsoft/onnxruntime,
microsoft/onnxruntime-openenclave,
onnx/onnx
and 4 other
repositories
Contribution activity
March 2021
Created 1 commit in 1 repository
Opened 1 pull request in 1 repository
microsoft/onnxruntime
1
merged
Reviewed 20 pull requests in 1 repository
microsoft/onnxruntime 20 pull requests
- Update CUDA custom op unit tests to account for recent ORT change
- Add support to save onnx graph with external initializers file.
- Update type reduction operator type usage processors set.
- Enable type reduction for Range, ReverseSequence, ScatterND, Split, and Unique CPU kernels.
- Release buffers for prepacked tensors
- [NNAPI EP] Possible fd leak in NNAPI NNMemory
- Enable type reduction in EyeLike, Mod, random.cc CPU kernels.
- Support parsing an array of values stored as an attribute in a custom op
- Enable build with bitcode for iOS
- Rename op kernel type control 'supported types' to 'default types'.
- [NNAPI EP] Add support of QlinearAveragePool
- Enable type reduction for Clip, MaxPool, and Pad CPU kernels.
- Support required types when excluding typed registrations
- [NNAPI EP] Add per-tensor u8s8 support for Qlinear[Conv/MatMul]
- [CoreML EP] Add [Average/Max]Pool support
- Support optional inputs/outputs in custom op development
- Change OpKernel class to be shared with shared providers
- Add support for op kernel type control required types, require int64 for some ops
- Refine force CPU fallback logic in the CUDA EP
- [CoreML EP] Add Concat support

