CHiME-5/6 speech datasets relicensed to CC BY 4.0 and now on Hugging Face
shinjiw_at_cmu · x · 2026-11-17
The CHiME Challenge team has relicensed the CHiME-5 and CHiME-6 speech datasets from CC BY-SA 4.0 to CC BY 4.0, making them easier to reuse in research and open model development.
- CHiME-5 contains recordings from 20 dinner parties held in real homes, targeting speech separation and recognition in everyday environments.
- The dataset is now hosted on Hugging Face (chimechallenge/chime5) with train/dev/eval archives, transcriptions, and floorplans, downloadable via hf download.
More from Research
- BitDecoding-FP4: native FP4 KV cache decoding via algorithm-system co-design — UoE-Mlsys · 2026-11-18
- HT-Policy: multimodality isn't why diffusion policies win at imitation learning — chris_j_paxton · 2026-11-18
- Hypothesis: Higher Adam β1 May Help More in Small-Batch, Low-LR Training Regimes — torchcompiled · 2026-11-18
- NeurIPS paper: Joint Energy Models show human vision aligns between generative and discriminative — ShahabBakht · 2026-11-18
- SWE-CoWork benchmark: best model fully passes only 8 of 101 handoff coding tasks — omarsar0 · 2026-11-18
- Pre-training trend watch: dropping LR decay for constant LR plus checkpoint averaging — rasbt · 2026-11-18