Guillaume Lagrange

Thanks for filing! We really need to update the notebook examples, or entirely remove them 😅 they haven't been updated in a while.

Sorry I haven't had the time to take a good look at this, I'll make time to give you some feedback at the beginning of next week 🙂

First of all, great work so far! It seems like you've come a long way 👏 Regarding the implementation, I went through the suggested changes in the PR a couple...

@McArthur-Alford Thank you for sharing the issues you ran into and your thought process! It helps to know why some decisions were made a bit more specifically. To give a...

> > This PR has been marked as stale because it has not been updated for over a month > > Apologies bot. I am in fact still alive, just...

> One thing I would like to support is different padding modes for when indices extend beyond the data, such as zero-padding, repeat-padding, and mirror-padding. In principle this choice would...

> I'm taking a stab at implementing this over at https://github.com/timstr/burn/tree/feat/grid_sample, so far I seem to have a working implementation of a `grid_sample_1d` op. I'll hopefully have 2D and 3D...

> I have implemented a custom backend that supports 4-dimensional padding with zero bilinear interpolation for grid sampling. Would you mind taking a look to see if there are any...

At a high-level, it could be added as a tensor operation similar to other utility methods in [`burn-tensor/src/tensor/grid`](https://github.com/tracel-ai/burn/tree/main/crates/burn-tensor/src/tensor/grid). But a naive implementation will not be the most performant. So it...

Where did copilot come from and why are we asking it reviews? 🤣 A little busy right now but will take some time to review this properly early next week!...