Woobin Park

Undergrad at Seoul National University

[Yilun Du @du_yilun]

Robotics has spent years scaling demonstration data and model size. What happens when we scale EXPLORATION for robot agents?

We're incredibly excited to share RAX, an agentic robotics system that learns by scaling online exploration.

A coding agent connects directly to the robot’s perception and control APIs, learns from feedback, and builds programs that perceive, plan, and act. The model stays frozen. As exploration scales, the agent improves its programs and builds up reusable skills.

The results:
�� ZERO teleoperation or action demonstrations on real robots—from insertion with 0.1 mm clearance and air-fryer manipulation to deformable objects and long-horizon mobile whole-body control.
• +37 percentage points on average over leading learned policies across five simulation benchmarks: RoboDojo, BEHAVIOR-1K, HumanoidArena, DexJoCo’s bimanual tasks, and LIBERO-Pro.
• One framework spanning single-arm, bimanual, dexterous, humanoid, and mobile robots, with new embodiments connected through APIs.

The craziest part: on each real-world embodiment, going from scratch to mastering the tasks took roughly two hours. No teleop. No model finetuning.

We’re open-sourcing the full framework AND every experiment session. We can’t wait to see what people build with it!

Project + sessions: https://t.co/O04Z5PUgqY
Paper: https://t.co/sd66qnYZWB
GitHub: N

Project fully led by @RunhanH and @xczhangxx who are absolutely awesome.

How it works, experiments, and robot videos below 🧵

(1/N)

[Woobin Park @parkwbin03]

@jayhxmo Thanks!

[Woobin Park @parkwbin03]

Excited to share our new work on one-step generation!
I personally think our work offers a complementary perspective on understanding generative modeling.

Huge thanks to my coauthor @Jaeyeon_Kim_0 and our advisors @du_yilun @sitanch Bohyung Han!

[Jaeyeon (Jay) Kim @Jaeyeon_Kim_0]

Today, we introduce Priming Models, a new paradigm for training one-step generative models from scratch.

No diffusion teacher. No flow map. No CFG. No JVP.

We believe this work will challenge how we think about building one-step models.

Joint work with @sitanch, @du_yilun, @parkwbin03, Bohyung Han.

🔥 1.19 FID one-step on ImageNet 256×256 while using substantially less compute, e.g., 12X less than Drifting-B, 6X less than Drifting-L
🎨 Scales to text-to-image generation (GenEval 0.82)
🤖 World modeling on RT-1 / Bridge-V2

Priming Model is built on one core message: [Reconstruction pretraining --> Distributional post-training].

[Sitan Chen @sitanch]

We figured out a much simpler method for training one-step generative models from scratch! Advances Pareto frontier for quality v FLOPs and strips out the engineering bells and whistles that made these so finnicky to train. Check out our codebase and try them out for yourself!

[Yilun Du @du_yilun]

Excited to share a very simple and effective recipe for one-step generation: simply do reconstruction pretraining → distributional post-training.

Priming Models achieve 1.19 FID on ImageNet 256×256 and achieve high-quality text-to-image generation and robot world modeling. Excited to share this work with @Jaeyeon_Kim_0 and the team!

[Jaeyeon (Jay) Kim @Jaeyeon_Kim_0]

Today, we introduce Priming Models, a new paradigm for training one-step generative models from scratch.

No diffusion teacher. No flow map. No CFG. No JVP.

We believe this work will challenge how we think about building one-step models.

Joint work with @sitanch, @du_yilun, @parkwbin03, Bohyung Han.

🔥 1.19 FID one-step on ImageNet 256×256 while using substantially less compute, e.g., 12X less than Drifting-B, 6X less than Drifting-L
🎨 Scales to text-to-image generation (GenEval 0.82)
🤖 World modeling on RT-1 / Bridge-V2

Priming Model is built on one core message: [Reconstruction pretraining --> Distributional post-training].