Grigory Malivenko

Update: I thought the problem was with dimension ordering (in previous model it was CHW), so, I've converted the model to other dimension ordering (HWC). A new error in a...

I've tried to use this library with npm, but also had similar problem: ``` These relative modules were not found: * ./texture_copy.glsl in ./~/keras-js/lib/Tensor.js * ./activation.glsl in ./~/keras-js/lib/ext/core/WebGLActivation.js * ./input_transform.glsl...

Hello @guoguoguilai. Can you provide me full source code so I can reproduce your error?

Hello @hyuan-kamuda. The project is in progress now. It supports basic layers such as Conv, Linear, some activations. You can try to calculate number of MACs using CLI (see more...

Hello @M-Tonin @vipulbjj . Unfortunately, recurrent layers are not supported now. ONNX produces a bunch of low-level operations, and something doesn't match somewhere in the middle. I don't want to...

Hello @Tian14267. Can you please share the model you try to convert?

Hello @teobaluta . Thank you for your PR. I'll check & merge hopefully this week. Can you also implement corresponding tests (located in `test` directory) ?

Hello @mdhimes! Thanks for your pull request, I'll review / test it as soon as possible.

Hello @SelComputas. I fixed problems related to Squeezenet (you can get latest changes from github, master branch). There was a problem with Dropout (in ONNX it returns not only operation...

Hello, @teobaluta . You can see how the code is being tested at `test/` . If it's impossible to test your code with PyTorch to ONNX conversion, you can use...