Huan Liu

This use case came up when I tried to train a model on colab, but the colab runtime got killed and I lost the wandb logs.

> @liuhuanjim013 Could you [run pre-commits](https://github.com/huggingface/lerobot/blob/main/CONTRIBUTING.md#submitting-a-pull-request-pr) to fix formating? LGTM otherwise thanks @aliberts ! done!

this is needed when visualizing dataset with episodes removed in the middle (https://github.com/huggingface/lerobot/pull/831)

Got the following combo working for Jetson Orin NX (lerobot still requires torchvision 0.21 though) - **JetPack 6.1 (L4T R36.4) + CUDA 12.6**: - [PyTorch 2.5](http://jetson.webredirect.org/jp6/cu126/+f/5cf/9ed17e35cb752/torch-2.5.0-cp310-cp310-linux_aarch64.whl) - [torchvision 0.20](http://jetson.webredirect.org/jp6/cu126/+f/5f9/67f920de3953f/torchvision-0.20.0-cp310-cp310-linux_aarch64.whl)

Thanks for doing this @bsprenger ! Worked great for me as well! I tried it and noticed a few things: 1. I removed an episode using your script 2. Checked...

+1 to the flag idea!

@bsprenger I tried it and checked the files online (I tried to remove episode 15-17 from a dataset of 50 episodes): 1. `episodes,jsonl`, `episodes_stats.jsonl`, `info.json` all seem to be updated...

@bsprenger tested on a different repo and it works! so it should be good for anyone to try the latest version! for the one that had trouble, i'll need to...

i tested on the old(broken) dataset, did the following to fix it: 1. removed (renamed) dataset on hugging face 2. used latest remove episode script to remove a new episode...

> one thing i noticed using the visualization script: when using down arrow key to see the next episode, i would get 500 Internal Server Error when navigating from the...