The interdisciplinary AI & Data Science center of Institut Polytechnique de Paris and HEC Paris
This is the open-source home of the Hi! PARIS engineering and research team. We build and release tools, models, benchmarks, and teaching materials across machine learning, deep learning, and applied AI — from foundational research to industry collaborations.
Follow this page to keep track of what we're working on. ⭐ Star the repos you find useful and feel free to open issues or contribute.
Neural text-to-speech, vocoders, prosody control, and voice conversion — with a focus on French and other European languages.
- Prosody-Control-French-TTS — End-to-end pipeline for French TTS with SSML prosody control (ICNLSP 2025)
- wavlm-vocoder-french — WavLM-to-audio neural vocoder for French speech reconstruction (JEP 2026)
- CosyVoice2-EU — Europeanized CosyVoice2 for French & German
- FakeParts — FakePartsBench: 25K+ AI-generated videos with pixel- and frame-level annotations of full and partial deepfakes
- deepdespeckling — Despeckling tools for SAR / radar imagery
- XPER — Measuring the contribution of features to the predictive performance of any econometric or ML model
- Tools for estimating and reducing the energy consumption of neural networks (NeuroEnergy, energy_estimator)
- Summer_Schools — Hands-on demos and notebooks from the Hi! PARIS Summer Schools (ML, Deep Learning, AI engineering)
- Hickathon — Materials from our flagship data science hackathon
- 🌐 Learn more at hi-paris.fr
- 🤗 Find our models and datasets on Hugging Face
- 📦 Browse all our repositories
- 🐛 Found a bug or have an idea? Open an issue on the relevant repo
- 🔭 Interested in collaborating? Reach out via our website
Built with ❤️ by the Hi! PARIS team — Institut Polytechnique de Paris · HEC Paris
