Dilshod Tadjibaev

The issue reported here has been resolved but there is another issue: ``` DEBUG burn_import::burn::graph: Registering node => 'concat' WARN onnx_ir::node::resize: Resize: coordinate_transformation_mode is ignored WARN onnx_ir::node::resize: Resize: cubic_coeff_a is...

This is a great idea. From the beginning I was intending Burn's ONNX IR to be its own crate. For now, however, we should keep it within the burn repo...

@BjornTheProgrammer FYI, it's sufficient to fix make this work with crates in this test: https://github.com/tracel-ai/burn/tree/main/crates/burn-no-std-tests

Tagging @nathanielsimard @laggui @syl20bnr to suggest dependency consistency. I will take a look closer tomorrow.

> This pull request only supports `ndarray`. Do you all think that exploration into other backend supports should be done, like `autodiff` or `candle`? We mainly target inference. We do...

> > Tagging @nathanielsimard @laggui @syl20bnr to suggest dependency consistency. > > I will take a look closer tomorrow. > > Any resolution on this front? Yes, the cubecl-common has...

> I've added documentation to the burn book (it should be rewritten more to your standards) and replaced the deprecated methods, I'm unsure if I did so correctly. One of...

@BjornTheProgrammer is this error you're gettin when building `run-checks.sh no-std`? ``` error: dependents require atomic CAS but not available on this target by default; consider enabling one of the `unsafe-assume-single-core`...

I resolved conflicts and rebase. Let me know what's blocking you.

@BjornTheProgrammer i will review in coming days. I also invited others to look into this. This will be a very useful feature especially with your example and a book section!