Extend ScatterElements op with min and max reduction
See original GitHub issueFollowing https://github.com/onnx/onnx/pull/3484, the ScatterElements operation supports add
and mul
reduction operations. Additionally supporting min
, max
, and mean
operations would allow for the remaining unsupported torch_scatter
ops to be supported by ONNX.
I think this would allow PyG models to be exported to ONNX. Currently the PyG softmax
implementation uses scatter_max
and scatter_sum
. I’m hoping that these changes would be fairly easy to implement since the other reduction operations are already supported.
Tagging @shubhambhokare1 who contributed https://github.com/onnx/onnx/pull/3484 for guidance.
Issue Analytics
- State:
- Created a year ago
- Comments:6 (1 by maintainers)
Top Results From Across the Web
ScatterElements - 16 vs 18 — ONNX 1.12.0 documentation
ScatterElements takes three inputs data, updates, and indices of the same ... (Opset 18 change): Adds max/min to the set of allowed reduction...
Read more >OpenMP min reduction and std::min - c++ - Stack Overflow
What the reduction clause implies is that each thread will have its own private copy initialized to the neutral element of the min...
Read more >Python Runtime for ONNX operators
Dn) before a BatchNormalization Op. This operator has optional inputs/outputs. ... The interval is specified by the inputs 'min' and 'max'.
Read more >Changelog - GitHub
The ONNX parser will automatically route ROIAlign ops through the plugin. ... factors to optimize softmax MAX reduction in versions 2 and 3...
Read more >Function-Level Support Status - Neural Network Libraries
Reduction. Arithmetic. Logical ... Expand. ✓. ✓. Reshape, Broadcast. EyeLike. X. Not yet implemented. Flatten ... Max, Neg, Constant, Min. LogSigmoid.
Read more >
Top Related Medium Post
No results found
Top Related StackOverflow Question
No results found
Troubleshoot Live Code
Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free
Top Related Reddit Thread
No results found
Top Related Hackernoon Post
No results found
Top Related Tweet
No results found
Top Related Dev.to Post
No results found
Top Related Hashnode Post
No results found
I have started a PR for it locally here, https://github.com/philass/onnx/pull/1
I will submit it to ONNX when it is more polished
@philass and I are looking in to this right now