Add tasks related to autoencoding and compression

Please consider adding tasks related to audio, image, video, and/or other types of lossy data compression.

There are many tasks related to computer vision and audio understanding, but none are suitable for papers that focus on compressing data.

My recommendation would be to add the following tasks:

Audio compression
Image compression
Video compression

Here are some popular models that are already built into the huggingface transformers/diffusers libraries focusing on compression:

https://huggingface.co/docs/transformers/model_doc/encodec
https://huggingface.co/docs/diffusers/api/models/autoencoder_dc
https://huggingface.co/docs/diffusers/api/models/autoencoderkl_cosmos