nicholas

This seems to work fine in vanilla vscode, just not with Dendron. Can we at least disable Dendron's handling of lists?

Thanks for your answer @vfdev-5. Using only a single GPU does work; I had already been training with a single GPU using `CUDA_VISIBLE_DEVICES` as you suggested. I had not tried...

Yes, that did work fine - as mentioned, I found and removed them so distributed training is working. I was wondering if it is possible to simply allow unused parameters?...

@mwip @emeryberger ## Fix There is a single line fix for the repro code above: use `persistent_workers=True` to the DataLoader at construction: ```python dl = torch.utils.data.DataLoader( ds, batch_size=8, num_workers=16, persistent_workers=True,...

Removing old resurrect files was also the solution for me, but that included these: ```bash rm -rf ~/.tmux/resurrect ``` Now auto restore works with these settings: ```bash set -g @plugin...

Any possibility of using a library like `nvtop`? https://github.com/Syllo/nvtop

Perhaps worth noting, in case it is used elsewhere: the `DirEntry` objects hold both the filenames as well as the full path - as two separate attributes. This version obviously...