Manthan Patel ✈️ CoRL

Vision foundation models like DINO and SigLip now power much of robot learning.

But they are built for RGB. What about depth?

Introducing DeFM: a foundation model for depth images, built for robotics.

🌐 de-fm.github.io

Weights, code, and demo are out today 🧵

Heading to #CoRL2026 in Austin 🤠 Come say hi!

📍 Wed 11 Nov: EgoHTR in Poster Session 3
📍 Thu 12 Nov, morning: DeFM at the Efficient Foundation Models for Real-Time Embodied AI workshop
📍 Thu 12 Nov, afternoon: WayTrace (spotlight), VLD, RangeNav, and WildOS at #SISREN

👇

We gave a robot the walking directions you'd get from a web map.

Then we let it find its own way through Zurich: crosswalks, stairs, crowded streets, construction sites, 2.35 km in total.

Introducing WayTrace 🧵👇

Getting a step closer to long-range city-scale Navigation powered by Foundation Models 🚀

A sneak peek into what we have been building. Stay tuned for the release

p.s. Can you guess this popular location?

Great to see people already starting to use our #EgoHTR dataset. We will be presenting this work at #CoRL2026.

Several motions, including backflips and ladder climbing 🪜🤸

Paper:
arxiv.org/abs/2607.13472
Dataset: huggingface.co/datasets/legge…