Principal Research Scientist @heyjasperai | prev. Research Scientist @StabilityAI | PhD in Machine Learning & Computer Vision from @Inria
Paris, France
Joined June 2020
- Tweets107
- Following186
- Followers315
- Likes437
Pinned Tweet
🔥🔥 Today we, @clipdropapp are launching our 💡image relighting #AI application 💡
The app allows you to apply professional lights to your portrait images 📸 in real time ⚡
Try it now! it is free 🙂
producthunt.com/posts/clipdr…
#photography #MachineLearning
Onur Tasar retweeted
Exciting news from Jasper Research! 🚀
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!
Our technical report shares everything we learned along the way.
👇
Onur Tasar retweeted
📢 New @heyjasper release ! 📢
MONET 🌸 : An Apache2.0 deduped and recaptioned dataset of 105M samples unlocking reproducible text-to-image research.
Nano T2I 🖌️ : A codebase to train your own T2I model
🤗 @huggingface: huggingface.co/datasets/jasp…
💻: github.com/gojasper/nano-t2i
Very excited about this new release, pushing the boundaries of open and reproducible T2I research.
Congrats to the team!
Benjamin Aubin Gonzalo Quintana @onurxtasar @UlaLaParis @_jeev2 @dh7net @clipdropapp @heyjasperai
Onur Tasar retweeted
We’re thrilled to announce that the Jasper Research team’s latest paper on Latent Bridge Matching (LBM) was accepted to the prestigious IEEE/CVF International Conference on Computer Vision! Thank you @ICCVConference for the recognition, and congratulations to the team: @onurxtasar, Sanjeev Sreetharan, Benjamin Aubin, Gonzalo Quintana, @dh7net, and @CChadebec.
LBM is a technique for rapid, high-fidelity image-to-image translation. Learn more about how LBM is changing visual marketing for retailers. bit.ly/3ZWMEcI
Onur Tasar retweeted
Proud to share the latest breakthrough from Jasper Research: Latent Bridge Matching (LBM) just got a major update!
Our team has released three new open-source models on Hugging Face—supporting depth, normals, and relighting—along with training code and a Gradio demo for object relighting.
This work continues our mission to advance controllable, high-quality AI image generation for enterprise marketing teams.
🔗 Explore the research, models, and demo: linkedin.com/feed/update/urn…
👏 Congrats to @CChadebec, Benjamin Aubin, @OnurXTasar, and @dh7net for pushing the field forward.
📢 Latent Bridge Matching update! 📢
New @heyjasperai release 🚀
- 3 LBM models on @huggingface (relighting, depth, normals)
- LBM training code
Also see our @Gradio space for object relighting!
🤗 huggingface.co/jasperai/LBM_…
💻 github.com/gojasper/LBM
📗 arxiv.org/abs/2503.07535
Onur Tasar retweeted
📢 Latent Bridge Matching update! 📢
New @heyjasperai release 🚀
- 3 LBM models on @huggingface (relighting, depth, normals)
- LBM training code
Also see our @Gradio space for object relighting!
🤗 huggingface.co/jasperai/LBM_…
💻 github.com/gojasper/LBM
📗 arxiv.org/abs/2503.07535
Onur Tasar retweeted
New research drop! Our team at Jasper is redefining image-to-image translation with LBM: Latent Bridge Matching—a scalable approach to high-speed AI-powered image transformation. See it in action!
Excited to share our new research at @heyjasperai !🚀
LBM: Latent Bridge Matching for Fast Image-to-Image Translation
Try out our @huggingface space for object relighting!
🤗@Gradio demo: huggingface.co/spaces/jasper…
👉Paper: arxiv.org/abs/2503.07535
💻Repo: github.com/gojasper/LBM
Thrilled to share our new research on image-to-image translation, demonstrating strong performance across tasks like object relighting, shadow prediction, object removal, and depth and surface normal estimation.
Check out our research paper, project page, demo, and Github repo👇
Excited to share our new research at @heyjasperai !🚀
LBM: Latent Bridge Matching for Fast Image-to-Image Translation
Try out our @huggingface space for object relighting!
🤗@Gradio demo: huggingface.co/spaces/jasper…
👉Paper: arxiv.org/abs/2503.07535
💻Repo: github.com/gojasper/LBM
Onur Tasar retweeted
We had an incredible time last Saturday at the prestigious International @RealAAAI Conference in Philadelphia! Jasper’s research team was proudly represented by Sanjeev Sreetharan and Onur Tasar, who presented our Flash Diffusion paper—a distillation method designed to boost the efficiency and speed of diffusion models in AI, particularly for image generation.
Huge congratulations to the entire Jasper Research team—@CChadebec, @OnurxTasar, Benjamin Aubin, and Eyal Benaroche—for this amazing achievement! 👏 Learn more about the research here: bit.ly/4gaFGXH
Our blog post introducing our new research on controllable shadow generation using single-step diffusion models trained on fully synthetic data👇
Creating realistic, high-impact product images is essential for connecting with customers—but achieving this at scale has always been a challenge. Fortunately, our Jasper Research Lab in Paris is changing that.
Our latest research and single-step shadow generation model give brands precise control over depth, softness, and intensity in product images, eliminating the need for costly photoshoots and time-consuming edits.
Our Research Scientist, @OnurXTasar, shares more about this breakthrough and how it will help reshape AI image generation in marketing: bit.ly/3Cqc75W
We (@heyjasperai / @clipdropapp research team) are excited to announce the release of our latest research project on fast and controllable shadow generation. Our 1-step diffusion model can create realistic shadows for object images in under a second ⚡️, while giving you precise control over shadow direction, softness, and intensity 🔨.
- Project page: gojasper.github.io/controlla…
- Research paper: arxiv.org/abs/2412.11972
- Demo : shadow-generation-demo.jaspe…
- Public test set on @huggingface : huggingface.co/datasets/jasp…
@CChadebec @benjamin_aubin_ @dh7net
During the training stage of the diffusion model, we incorporated polar and azimuthal angles along with the light size parameter into the denoiser using vector conditioning. This conditioning mechanism allows precise control over the shadow direction and softness. By training the diffusion model with rectified flow, we achieved the ability to generate shadows in just a single step. The model generates the shadow map as an alpha layer, allowing it to easily integrate with any background image. The full workflow is detailed below.👇
In the absence of a public benchmark on the topic of controllable shadow generation, we created a synthetic test set with three distinct tracks: horizontal direction control, vertical direction control, and softness control. The test set is publicly available on @huggingface 🤗
huggingface.co/datasets/jasp…
Onur Tasar retweeted
Proud moment for Jasper Research: “Flash Diffusion,” our paper on a new method designed to enhance the efficiency and speed of diffusion models in AI for image generation, has been accepted to @RealAAAI .
👏Congrats @CChadebec @onurxtasar Eyal Benaroche @benjamin_aubin_ for this incredible achievement! bit.ly/4gaFGXH
Onur Tasar retweeted
🎉 Very happy to share that @heyjasperai's very first research paper, Flash Diffusion just got accepted to #AAAI.
This paper introduces a novel and flexible distillation method designed to speed up image generation with diffusion models.
👉 Paper: arxiv.org/abs/2406.02347
Onur Tasar retweeted
🚀 Excited to launch Texture Diffusion, an open-source Blender add-on that brings the power of diffusion models to texture generation! 🎨
✨ Features:
🎯 Inpainting for precise texture edits
🔗 Support for LoRAs & IP-Adapters
🤝 Seamless integration into Blender’s workflow
github.com/Shaamallow/textur…
Inspired by Stable ProjectorZ, this is a small project I’m happy to share with the community! Feedback & suggestions are welcome!
The add-on is making use of the popular @ComfyUI so that you don't have to install yet another Diffusion Backend, and @cubiq IP-Adapters integration into #ComfyUI.
⭐ Check it out, star it on GitHub, and share it with friends if you like it!
#blender #ai #diffusion #opensource
Onur Tasar retweeted
Our team at Jasper Research @heyjasperai has just released Style-Rank 🎨, a fair and unified benchmark of open-source training-free styling methods.
Our evaluation is performed on our Style-Rank evaluation set, compiled from multiple styling papers and provided on @huggingface. You can check our evaluation code with the links below
📽 Project page: gojasper.github.io/style-ran…
⌨️ Codebase: github.com/gojasper/style-ra…
🤗 Dataset: huggingface.co/datasets/jasp…
@benjamin_aubin_, @cchadebec, @onurxtasar, @dh7net
Onur Tasar retweeted
New @heyjasperai research release 🚀
Upscaler, depth and normal maps ControlNets for FLUX.1-dev by @bfl_ml
@huggingface Models 🤗:
- Upscaler: huggingface.co/jasperai/Flux…
- Normals: huggingface.co/jasperai/Flux…
- Depth: huggingface.co/jasperai/Flux…
@Gradio Demo:
huggingface.co/spaces/jasper…