Juan Emmanuel Johnson
Juan Emmanuel Johnson
Hello, Is this project still under the closed-community restrictions?
Hi @LysSanzMoreta, I have an example of how one could do SparseGPs (the classic variational free energy version). You can find the demo script [here](https://github.com/jejjohnson/gp_model_zoo/blob/master/code/numpyro/src/sparse.py) as well as the jupyter...
Hi, I just wanted to know if this was being worked or will it be worked on in the near future? I see the repo referenced has been archived so...
Hi @fritzo, thank you for the information! A side question: the `flowtorch` library will be for PyTorch (at least it was the last time I checked). Are there plans to...
Thanks for the advice @patrick-kidger ! I will do the same. I'm also more familiar with the PyTorch dataloaders as well. Do you do similar practices to, for example, what's...
> A word of caution: do not make your `collate` return `np.array` instead of `Tensor` (and I think the same applies to return values from the dataset class). It's convenient...
Hi thanks for you interest!  So in order to debug, would you be able to provide a minimal working code example? Then maybe I could see if there is...
Good point, this is specifically for the viz team. In principle, they can easily query the same images that we download/have for the ML team. But there might be some...
Good point. Changed the tags and updated the assignees. If there's any extra preprocessing we can do, they can tag us.
@Lkruitwagen No, this is for the far future. It will be an extended feature that is MLOPs dependent.