README / README.md
jonbarker68's picture
Update README.md
6370b5d verified
|
Raw History Blame Contribute Delete
1.75 kB
---
title: README
emoji: 🎤
colorFrom: blue
colorTo: gray
sdk: static
pinned: false
---
# CHiME Challenge
**Speech separation and recognition in real, everyday acoustic environments.**
The CHiME challenge series releases conversational, multi-channel and wearable speech recordings made in the places people actually talk — homes, cars, cafés, streets, and more recently through smart glasses — and runs open evaluations on them. The emphasis throughout has been on real recordings rather than simulation, and on tasks that stay close to how the audio would really be captured and used.
This organization hosts the openly licensed CHiME datasets.
## Datasets
- [**chime3-background-noise**](https://huggingface.co/datasets/chimechallenge/chime3-background-noise) — CHiME-3 multichannel background noise recordings (BUS / CAF / PED / STR), six-channel tablet array, 16 kHz. CC BY 4.0. Mirrors the [Zenodo record](https://doi.org/10.5281/zenodo.23571515), which is the one to cite.
- [**chime5**](https://huggingface.co/datasets/chimechallenge/chime5) — CHiME-5 speech separation and recognition dataset with recordings from 20 dinner parties in 20 homes. Includes training, development and evaluation recordings, transcriptions, and floor plans. CC BY 4.0.
- [**chime6**](https://huggingface.co/datasets/chimechallenge/chime6) — CHiME-6 speech separation and recognition dataset. Supersedes CHiME-5 with clock-drift-corrected audio across all 20 dinner parties, updated transcriptions, and official splits used in CHiME-6 and CHiME-7 DASR. CC BY 4.0.
## Links
- Challenge website and full history of past challenges: <https://www.chimechallenge.org/>
- Current edition: [CHiME-10](https://www.chimechallenge.org/challenges/chime10/)