Dilshod Tadjibaev

@ShivangRawat30, I would recommend you start with a unary operator Sqrt. Work on this and submit a PR. If you agree, just comment claim Sqrt. If you have questions and...

@AuruTus, since all these ops already available in Burn, they have to be implemented in burn-import. I would recommend to start with something simple. Maybe with Tanh?

@AuruTus we added onnx file base testing to verify end to end conversion. Let me know if you have questions.

@CohenAriel I am not sure what it is being compared to. Is it expected code or generated code?

In that can you will need to modify your expected test. You may have to use custom instead of using the macros in place. You should find some examples with...

@nathanielsimard @louisfd > Regarding the linear conversion: I'm unsure we should "force" our modules to store the weights the same way as PyTorch. Yes, I agree it's not ideal, as...

@nathanielsimard and I had an offline conversation, and here is the revised summary: 1. We agreed that the primary goal is to effectively integrate PyTorch weights into the Burn framework...

Just to update everyone. I have a solution that will accomplish what @nathanielsimard and I discussed. I researched serde extensively and it's possible to achieve only through a custom deserializer....

> Fine for me too! Thanks a lot for your hard work! 😃 > > Just a final question, the `TODO` in the first message will be covered in a...

@carrotflakes it has been a month and we don't we any open PR related. You can give a try. We will review your quickly.