Bardienus Duisterhof
@bardienus • 1,478 subscribers
Incoming Postdoc @stanford || PhD Student @CMU_Robotics with @jeff_ichnowski || Efficient World Models for Robot Manipulation 📷 ||
Shorts
Videos
Introducing PointZero—a 3D world model pre-trained without robots. Dexterous manipulation requires understanding diverse 3D dynamics, but current approaches rely on expensive robot data. So how can we pre-train a 3D dynamics model without it? PointZero introduces a simple idea: learning to complete 3D point tracks yields transferable 3D dynamics! Our model achieves SOTA results on: 🔥 Zero-shot 3D dynamics 🔥 Action-conditioned 3D dynamics (post-training) 🔥 Imitation learning (post-training) Details and links 👇 1/6
Bardienus Duisterhof29,119 просмотров • 1 месяц назад
![Introducing Modality Forcing, a recipe for post-training T2I models for SOTA RGB-Depth generation! Text-to-image (T2I) models learn rich representations of the spatial world. How do we build on this prior for high-quality depth generation? 🧵 [1/6]](https://image.24vids.com/tw-2065437358829531237/amplify_video_thumb/2065436426020478976/img/5jxAY5zzuttB0qSW.jpg)
Introducing Modality Forcing, a recipe for post-training T2I models for SOTA RGB-Depth generation! Text-to-image (T2I) models learn rich representations of the spatial world. How do we build on this prior for high-quality depth generation? 🧵 [1/6]
Bardienus Duisterhof67,032 просмотров • 4 месяцев назад
Больше нет контента для загрузки