huggingface / huggingface/datasets

Mosaic Streaming (MDS) Support

Open
#6,736 1 comment 1 reaction 0 assignees View on GitHub
enhancement
Dominant language
Python
Stars
22k
Forks
3.4k
Avg merge
5d 7h
Merged PRs (30d)
17

Description

### Feature request

I'm a huge fan of the current HF Datasets `webdataset` integration (especially the built-in streaming support). However, I'd love to upload some robotics and multimodal datasets I've processed for use with [Mosaic Streaming](https://docs.mosaicml.com/projects/streaming/en/stable/), specifically their [MDS Format](https://docs.mosaicml.com/projects/streaming/en/stable/fundamentals/dataset_format.html#mds).

Because the shard files have similar semantics to WebDataset, I'm hoping that adding such support won't be too much trouble?

### Motivation

One of the downsides with WebDataset is a lack of out-of-the-box determinism (especially for large-scale training and reproducibility), easy job resumption, and the ability to quickly debug / visualize individual examples.

Mosaic Streaming provides a [great interface for this out of the box](https://docs.mosaicml.com/projects/streaming/en/stable/#key-features), so I'd love to see it supported in HF Datasets.

### Your contribution

Happy to help test things / provide example data. Can potentially submit a PR if maintainers could point me to the necessary WebDataset logic / steps for adding a new streaming format!

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.