4.3.0: group ONNX PRs

Dmitry Kurtaev 2020-03-28 21:59:59 +03:00
parent 3317793d2a
commit 5046bdcdf0

@ -20,13 +20,7 @@ Highlights of this release:
- Improved layers / activations / supported more models:
- LSTM from ONNX: [#16817](https://github.com/opencv/opencv/pull/16817)
- Broadcasting from ONNX: [#16738](https://github.com/opencv/opencv/pull/16738)
- Div op for constants: [#16722](https://github.com/opencv/opencv/pull/16722)
- Slice layer with multiple inputs: [#16715](https://github.com/opencv/opencv/pull/16715)
- ONNX: [LSTM](https://github.com/opencv/opencv/pull/16817), [Broadcasting](https://github.com/opencv/opencv/pull/16738), [Algebra over constants](https://github.com/opencv/opencv/pull/16722), [Slice with multiple inputs](https://github.com/opencv/opencv/pull/16715)
- DarkNet: grouped convolutions, sigmoid, swish, scale_channels