rustrust

see https://github.com/SergioBenitez/Rocket/issues/1238, the autobahn websocket test suite should be included as part of this effort

f_linalg_matrix_rank does Err on this, which is nice. The error message, however, is a page of text and doesn't include the matrix rank. I will file an issue with pytorch...

I am encountering this same problem when I am calculating only SVD via tch::Tensor::svd using: ``` let (u, s, v) = my_tensor.reshape(&[1, 1, 1024, 1024]).svd(true, true); ``` The PyTorch folks...

Start with the mnist/mnist_conv.rs file, copy it into a new example which we can refer to as mnist_conv_svdbug.rs You may need a little bit of patience to reproduce this bug....

I think this is working for me. For others who come across this, I used the custom-optimizer as an example of how to proceed here and did find it sufficient....

does h2spec not test for this...?

Would it be possible to have cookies as a first-class feature in ehttp?

I see set-cookie in the response headers when visiting common websites. Yes I am asking about proper overall handling of the cookie including storage for both native and wasm egui...