Our great sponsors
-
ndarray
ndarray: an N-dimensional array with array views, multidimensional slicing, and efficient operations (by RustyBamboo)
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
I created clblast-rs, which binds to the clblast library. Main reason to use it is if you want something relatively fast that also allows you to write your own kernels to operate on the same buffers, that clblast uses.
Although very primitive (only with element-wise operations and dot product operations), and probably would not fit your criteria of "good" crate, you can checkout my experimental fork of the ndarray crate with WGPU support: https://github.com/RustyBamboo/ndarray/
An example of it's usage: https://github.com/RustyBamboo/rust-grad
Interesting! I'm also working on an autograd framework based on ndarray, it would be cool, if you are interested or have time, to try and experiment with GPUs!