# Scaling robotics datasets with video encoding

DevFeed: [Scaling robotics datasets with video encoding](<https://devfeed.tech/articles/scaling-robotics-datasets-with-video-encoding-7554.md>)

Original publisher: [Read original article](<https://huggingface.co/blog/video-encoding>)

Author: Simon Alibert; Remi Cadene; Miquel Farré

Published: 2024-08-27T00:00:00Z

Content type: article

Language: en

Sources: [Hugging Face - Blog](<https://devfeed.tech/sources/hugging-face-blog.md>)

Topics: [robot sense of touch](<https://devfeed.tech/topics/robot-sense-of-touch.md>), [datasets](<https://devfeed.tech/topics/datasets.md>), [Compression](<https://devfeed.tech/topics/compression.md>), [data](<https://devfeed.tech/topics/data.md>)

Tags: [compression](<https://devfeed.tech/tags/compression.md>), [data](<https://devfeed.tech/tags/data.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [images](<https://devfeed.tech/tags/images.md>), [jpeg](<https://devfeed.tech/tags/jpeg.md>), [robotics](<https://devfeed.tech/tags/robotics.md>), [tools](<https://devfeed.tech/tags/tools.md>), [training](<https://devfeed.tech/tags/training.md>), [video](<https://devfeed.tech/tags/video.md>), [visualization](<https://devfeed.tech/tags/visualization.md>)

## AI overview

The article presents video encoding as a way to store the visual modality of robotics datasets more efficiently. It reports substantial size reductions while preserving training quality, with comparable single-frame loading and faster decoding for multiple successive frames, and describes tools for browsing and visualizing the datasets.

## Source excerpt

We're on a journey to advance and democratize artificial intelligence through open source and open science.