akern40

Thanks for the responses, I figured that might be the case. > 1. Extend the model PI to allow to (optionally) pass some information about the transforms down to the...

Ah ya I like challenges but I'm not a sadist. I'll wait until that PR is complete, and I can take the time to consider some design options to propose...

> As far as implementing the RFC goes, I'd be happy for someone else to implement it. Hi! I hope this isn't presumptuous, but I've started an implementation of this...

Ok, I've made some progress on the implementation and (in doing so) gotten a much stronger understanding of `ndarray`'s guts. I've run into a challenge, however, that I'm curious to...

Hey @grothesque! Glad you're interested, thanks for the comments. I saw you also commented on the maintainer issue - I'm just gonna keep this thread about this particular RFC and...

I've been wanting to get involved in `ndarray` development, and this seemed like a good place to get my feet wet! Hope you don't mind, but I've put in a...

> I suspect that the very high level of complexity of this library plays a role in limiting contributions. And my gut feeling is that a fair share of this...

Thanks for the feedback, @bluss! Can you chime in at all about your opinions on keeping the data-level `Arc` and `Cow`? It feels to me like that's a big design...

> It's funny I stumbled into a [use case](https://github.com/tracel-ai/burn/blob/main/crates/burn-ndarray/src/tensor.rs) today, the same day I am reading your comment. @daniellga funny how those things happen! Question about that usage: is it...

> I would be happy to add you to the team. > > My responsibility is I think - still - to get 0.16.0 out (with more or less the...