AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
WildOS: Open-Vocabulary Object Search in the Wild
DeFM: Learning Foundation Representations from Depth for Robotics
DeFM: depth foundation models for robotics. ViT and CNN encoders pretrained on 60M depth images, plus classification and segmentation benchmarks.
-
leggedrobotics/defm
Image Feature Extraction • Updated • 1.34k • 11 -
DeFM Feature PCA
🧊1Visualize DeFM depth foundation model features with PCA
-
DeFM: Learning Foundation Representations from Depth for Robotics
Paper • 2601.18923 • Published • 4 -
leggedrobotics/defm-bench-tartanground
Viewer • Updated • 1.2k • 36
DeFM: depth foundation models for robotics. ViT and CNN encoders pretrained on 60M depth images, plus classification and segmentation benchmarks.
-
leggedrobotics/defm
Image Feature Extraction • Updated • 1.34k • 11 -
DeFM Feature PCA
🧊1Visualize DeFM depth foundation model features with PCA
-
DeFM: Learning Foundation Representations from Depth for Robotics
Paper • 2601.18923 • Published • 4 -
leggedrobotics/defm-bench-tartanground
Viewer • Updated • 1.2k • 36
A novel VQA benchmark for VLMs that evaluates models on their embodiment-specific understanding of navigation across challenging real-world scenarios.