Sugato Ray

@awni Is there any `mlx.nn` equivalent of `torch.nansum` (if `np.nansum` is to be avoided)?

Thanks, I looked at the torch.nansum implementation as well. https://github.com/pytorch/pytorch/blob/014f91a9d9f94ac9a7f0711600240d7cd7f69844/torch/_decomp/decompositions.py#L4277-L4278 ```py def nansum(x: mx.array, axis: int=-1): return mx.sum(mx.where(mx.isnan(x), 0, x), axis=axis) mx.nansum = nansum ``` @awni Can we add this...

@awni What should I use as a substitution for `torch.tensor.median()`? - is there anything equivalent in `mx.array`? - can I instead use `np.median(x, axis=1)` as a placeholder for `torch.median(x, dim=-1)`?

Has this been resolved? I tried with my macbook pro m1, and am getting the same error. I added the path to `PATH` variable in addition to declaring the ffmpeg-executable...

I have tried with both brew and direct installation via Chronicler. But have not been able to make it work yet. Chronicler stays stuck as `Chronicler stopping...` forever, when you...

Seeing the same issue on `Ubuntu 20.04` with VS Code 1.60.2. @ShyykoSerhiy Am I missing something? Or, this is a bug (unfixed yet)?

@dvsrepo Of course, I can help. But it would be nice to know who from the repository core team will be mostly maintaining it as newer versions are released on...

@frascuchon Great. We will need a new conda-forge package for argila (the old rubrix package remains as is). Let's talk about this after this weekend.

@frascuchon Is there anyone from the core team who has had some experience with conda-forge packaging?

@frascuchon Could you please identify someone from the core Argilla team who will be a co-maintainer of the new feedstock repository for conda-forge presence of Argilla?