TTT3R: Enhancing 3D Reconstruction via Test-Time Training

The ICLR 2026 paper TTT3R introduces a test-time training approach with a simple state update rule to enhance 3D reconstruction. It effectively improves the length generalization of RNN-based models when handling inputs beyond their training context length.

2026-07-31 ~ 2026-07-31 · 2 related posts

• ICLR 2026 Paper TTT3R: Enhancing 3D Reconstruction via Test-Time Training — rsasaki0109 · 2026-07-31
• TTT3R: Test-Time Training Enhances Length Generalization in 3D Reconstruction — rsasaki0109 · 2026-07-31