Gen DGRL Dataset[Feature Request]

[braj29]

Motivation

Want to add the Gen DGRL Dataset for benchmarking RL algorithms in a easy & efficient way

Solution

Use D4RL wrapper (https://github.com/Farama-Foundation/D4RL) to download and load the dataset ready for training using simple python commands

Alternatives

For larger datasets it maybe best to download the files partially and then delete them as training progressing while simultaneously downloading remaining data

Additional context

Add any other context or screenshots about the feature request here.

Checklist

• [x] I have had an initial discussion with @vmoens
• [ ] Get familiar with TorchRL and GenDGRL Dataset
• [ ] Create a wrapper to download and load the dataset

[vmoens]

Here too, I'd rather not use anything D4RL-based. The library is 100% unmaintained, and it's a disaster to install. I believe we have everything we need in-house to build this. Have a look at how torchvision does the data download and extraction, and feel free to copy paste parts of their code if needed! https://github.com/pytorch/vision/tree/main/torchvision/datasets

[braj29]

Thanks a ton @vmoens ! I will take a look at how they have used torchvision datasets and do it that way

[vmoens]

Closed by #1773

[sr1998]

I have had problems uncompressing the dataset files and have not only used torchrl but also some custom methods of doing it. The error I get is: ReadError: not an lzma file

After manually extracting the files, it seems to not be possible to read the files and the content of the files seem to be inaccessible.

OS: Windows