Clément Chadebec

Research Scientist at Jasper AI

I am a research scientist at Jasper AI working on diffusion, flow and bridge models for image generation as well as distillation methods to make them run faster. I also lead the development and release of open-source projects. I have a Ph.D in applied mathematics and machine learning focusing on generative models at INRIA. In particular, I worked on Variational Autoencoders (VAE) and was mostly interested in the modelling of their latent space. I also maintain Python libraries trying to democratize those models. Find more on github.

News

Sep 28, 2026 I am joining Runway as a Member of Technical Staff, working on world models.
Sep 2, 2026 We built an internal 4B text-to-image model from scratch that turned out to be one of the fastest models out there (1k image generation in less than a second 🏎️). See our technical report that shares everything we learned along the way!
May 20, 2026 We release the MONET dataset, a 100M+ text-to-image dataset meticulously filtered, deduplicated, and multi-VLM re-captioned. Check out our Paper and the Dataset!
Jun 25, 2025 Our paper LBM (Latent Bridge Matching) is accepted to ICCV 2025 as a Highlight! 🎉
Mar 14, 2025 We release LBM: Latent Bridge matching for fast image-to-image translation! Check the paper out on arXiv and demo for image relighting on Hugging Face.
Dec 10, 2024 Flash Diffusion is Accepted to AAAI 2025 as an Oral Presentation! ⚡
Sep 26, 2024 We release three new Flux ControlNets (Image upscaling, depth and normals)! Check them out on Hugging Face! 🤗
Jul 3, 2024 We release a new demo combining Flash Diffusion and existing LoRAs without any fine-tuning! Check it out on Hugging Face! 🤗
Feb 22, 2024 I am now joining Jasper AI as a research scientist after the acquisition of Clipdrop. Let’s push the boundaries of Diffusion Models! 🚀
Oct 4, 2023 I am joining Stability AI as a research scientist. Let’s push the capabilities of generative models 🚀