Daniel Falbel

This is unfortunatelly not a simple task as LibTorch doesn't provide pre-built binaries for aarch64 systems, so you will need to first: 1. Build LibTorch from source (see instructions [here](https://github.com/pytorch/pytorch/blob/main/docs/libtorch.rst#building-libtorch-using-cmake))....

Sorry, I meant pointing to the correct LibTorch version (currently it's 2.0.1), when building LibTorch from source: https://github.com/mlverse/torch/blob/fe5020f62e6aa6c3c7fbce06c4aacb1e09c55db7/src/lantern/CMakeLists.txt#L4

Yes, you have to clone from the tag v2.0.1 in order to get the correct LibTorch version.

The hover to see full text is not implemented yet indeed. I think for verification, you can try something like @wesm described in: https://github.com/posit-dev/positron/pull/4121#issue-2423980667 The UI should keep responsive, whereas...

When torch creates a parallel dataloader (num_workers > 1) it will create some new R processes using callr and then copy the dataset you passed on into each one of...

Hi @kv9898 ! Thanks for trying it out and for the bug report. I think it's worth opening a separate issue with more details and what behavior you'd have expected....

Hi @MaximilianPi , I believe this is expected unfortunatelly. When building autoregressive models, since you have tensors that `requires_grad` in the computation, torch is storing the full computation graph in...

I recently updated this to `type="binary"` unaware of this error on Linux. If you do `install.packages("torch")` after setting up the repository, you should get the pre-built binaries anyway in all...

I think that would be nice, although they shouldn't change much. I don't remember why I did this change https://github.com/mlverse/torch/commit/547b5640839e0cbe1fb32863d6ef3ae69bf7724d unfortunatelly.

Thanks for the suggestion! Do you think it should be a user level configuration or users would prefer different behaviors depending on the dataset they are opening?