iiinova — where your digital self keeps evolving

Reposted by Lucas Degeorge

David Picard @davidpicard.eurosky.social · 03/09/2026

It's a cool paper, both old school and new school. There's an Easter egg: I designed a game for the teaser. Did you get it right? I tried it in my ML course and almost nobody got it.

3194

Lucas Degeorge @lucasdegeorge.bsky.social · 03/09/2026

🚀 New paper: Balancing Frequencies and Pixels in Flow Matching We tackle the low-frequency bias in pixel-space flow matching and train JiT up to 40% faster without any architectural changes. 📄 Read it here: arxiv.org/abs/2609.02748

1248

Reposted by Lucas Degeorge

David Picard @davidpicard.eurosky.social · 08/04/2026

🚨 arxiv.org/abs/2604.06129 PoM: A Linear-Time Replacement for Attention with the Polynomial Mixer This paper is the result of doing a lab-wide hackathon on an idea I've had for some time. Probably the paper with the highest number of authors I've ever done. It's a CVPR Findings 26. Thread 🧵👇

arxiv.org

PoM: A Linear-Time Replacement for Attention with the Polynomial Mixer

This paper introduces the Polynomial Mixer (PoM), a novel token mixing mechanism with linear complexity that serves as a drop-in replacement for self-attention. PoM aggregates input tokens into a comp...

46520

Reposted by Lucas Degeorge

David Picard @davidpicard.eurosky.social · 15/05/2026

👏 Folks! If you are curious about the Generative Modeling via Drifting paper, but you find it difficult to understand → I wrote a different interpretation of it. It's called: "An Expectation-Maximization interpretation of Generative Modeling via Drifting" davidpicard.github.io/pdf/An_Expec...

2285

Lucas Degeorge @lucasdegeorge.bsky.social · 21/05/2026

🎉 Our work MIRO is accepted to #ICML2026 @icmlconf.bsky.social We integrate human preferences directly during pretraining with multi-reward conditioning. ⚡MIRO is 19x faster than baselines and 370x cheaper at inference! 🤗 Try out the models: huggingface.co/spaces/nicol... See you in Seoul 🇰🇷 !

huggingface.co

MIRO - a Hugging Face Space by nicolas-dufour

Multi-reward conditioned text-to-image diffusion (ICML 2026)

1114

Reposted by Lucas Degeorge

Nicolas Dufour @nicolasdufour.bsky.social · 20/05/2026

Thrilled to share that MIRO is accepted to ICML 2026 @icmlconf.bsky.social ! 🎉 By training on the reward scores, we can simply condition the model on high rewards at inference time to guarantee top-tier, aligned outputs. We’ve updated our paper with some additional results!

arxiv.org

MIRO: MultI-Reward cOnditioned pretraining improves T2I quality and efficiency

The default paradigm of post-training text-to-image generators includes post-hoc selection of generated images, and subsequent training with one reward model to align the generator to the reward, typi...

1366

Lucas Degeorge @lucasdegeorge.bsky.social · 31/10/2025

Check out our new work: MIRO No more post-training alignment! We integrate human alignment right from the start, during pretraining! Results: ✨ 19x faster convergence ⚡ ✨ 370x less compute 💻 🔗 Explore the project: nicolas-dufour.github.io/miro/

nicolas-dufour.github.io

MIRO: Multi-Reward Conditioning for Efficient Text-to-Image Generation

Train once, align many rewards. MIRO achieves 19× faster convergence and 370× less compute than FLUX while reaching GenEval score of 75. Controllable trade-offs at inference time.

093

Reposted by Lucas Degeorge

Nicolas Dufour @nicolasdufour.bsky.social · 31/10/2025

We introduce MIRO: a new paradigm for T2I model alignment integrating reward conditioning into pretraining, eliminating the need for separate fine-tuning/RL stages. This single-stage approach offers unprecedented efficiency and control. - 19x faster convergence ⚡ - 370x less FLOPS than FLUX-dev 📉

37116

Reposted by Lucas Degeorge

David Picard @davidpicard.eurosky.social · 31/10/2025

👀 arxiv.org/abs/2510.25897 Thread with all details coming soon!

1132

Reposted by Lucas Degeorge

Vicky Kalogeiton @vickykalogeiton.bsky.social · 08/10/2025

Very proud of our recent work, kudos to the team! Read @davidpicard.bsky.social’s excellent post for more details or the paper arxiv.org/pdf/2502.21318

arxiv.org

0176

Reposted by Lucas Degeorge

David Picard @davidpicard.eurosky.social · 08/10/2025

Final note: I'm (we're) tempted to organize a challenge on that topic as a workshop at a CV conf. ImageNet is the only source of images allowed and then you compete to get the bold numbers. Do you think there would be people in for that? Do you think it would make for a nice competition?

284

Reposted by Lucas Degeorge

David Picard @davidpicard.eurosky.social · 08/10/2025

🚨Updated: "How far can we go with ImageNet for Text-to-Image generation?" TL;DR: train a text2image model from scratch on ImageNet only and beat SDXL. Paper, code, data available! Reproducible science FTW! 🧵👇 📜 arxiv.org/abs/2502.21318 💻 github.com/lucasdegeorg... 💽 huggingface.co/arijitghosh/...

14410

Reposted by Lucas Degeorge

Nicolas Dufour @nicolasdufour.bsky.social · 31/07/2025

I had the privilege to be invited to speak about our work "Around the World in 80 Timesteps" at the French Podcast Underscore! If you speak french, i highly recommend it they did a great job with the montage! If you want to learn more nicolas-dufour.github.io/plonk www.youtube.com/watch?v=s5oH...

youtube.com

Il a conçu la première IA d’OSINT (terrifiant… et génial)

YouTube video by Underscore_

0155

Reposted by Lucas Degeorge

David Picard @davidpicard.eurosky.social · 25/06/2025

If you want to listen to Nicolas (in French) talking about generative models for geolocation, it's right now: m.twitch.tv/micode

Lucas Degeorge @lucasdegeorge.bsky.social · 05/03/2025

🚨 News! 🚨 We have released the models from our latest paper "How far can we go with ImageNet for text-to-image generation?" Check out the models on HuggingFace: 🤗 huggingface.co/Lucasdegeorg... 📜 arxiv.org/abs/2502.21318

huggingface.co

Lucasdegeorge/CAD-I · Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

1153

Reposted by Lucas Degeorge

Vicky Kalogeiton @vickykalogeiton.bsky.social · 04/03/2025

Text-to-image models are trained on billions of data. But, is it necessary? Our "How far can we go with ImageNet for T2I generation?‬" @lucasdegeorge.bsky.social @arrijitghosh.bsky.social @nicolasdufour.bsky.social @davidpicard.bsky.social shows that no, if we are careful arxiv.org/abs/2502.21318

0194

Reposted by Lucas Degeorge

Mickael Chen @mickaelchen.bsky.social · 03/03/2025

Wow, neet! Reannotation is key here. Conjecture: As we are get more and more well-aligned text-image data, it will become easier and easier to train models. This will allow us to explore both more streamlined and more exotic training recipes. More signals that exciting times are coming!

122

Reposted by Lucas Degeorge

Shyamgopal Karthik @shyamgopal.bsky.social · 03/03/2025

These are some ridiculously good results from training tiny T2I models purely on ImageNet! It's almost too good to be true. Do check it out!

032

Reposted by Lucas Degeorge

David Picard @davidpicard.eurosky.social · 03/03/2025

🚨 New preprint! How far can we go with ImageNet for Text-to-Image generation? w. @arrijitghosh.bsky.social @lucasdegeorge.bsky.social @nicolasdufour.bsky.social @vickykalogeiton.bsky.social TL;DR: Train a text-to-image model using 1000 less data in 200 GPU hrs! 📜https://arxiv.org/abs/2502.21318 🧵👇

26516

Reposted by Lucas Degeorge

Nicolas Dufour @nicolasdufour.bsky.social · 10/12/2024

🌍 Guessing where an image was taken is a hard, and often ambiguous problem. Introducing diffusion-based geolocation—we predict global locations by refining random guesses into trajectories across the Earth's surface! 🗺️ Paper, code, and demo: nicolas-dufour.github.io/plonk

89732