Show HN: Open dataset for distilling GPU audio2face models to CPU students

Category: other

Tags: dataset, emotion-recognition, affective-computing

Score: 4.8/10 (Innovation: 3, Technical: 5, Documentation: 6, Utility: 5)

This project provides the CREMA-D dataset, a collection of 7,442 audiovisual clips of actors expressing six emotions, useful for research in emotion recognition and affective computing. It includes crowd-sourced ratings and detailed metadata, making it valuable for training multimodal emotion detection models. The project is a well-established academic dataset but lacks novel technical innovation or broad utility beyond research settings.

Target audience: data scientists, researchers, ai-ml engineers

Repository: https://huggingface.co/datasets/myned-ai/audio2face-emotion-arkit-teacher · R · NOASSERTION · 534 stars

View on Hacker News