FreeFlow: A Bias-free Hierarchical Transformer for Optical Flow Estimation
FreeFlow is a bias-free hierarchical transformer achieving state-of-the-art optical flow results on Sintel, KITTI-2015, and Spring benchmarks.
FreeFlow replaces task-specific inductive biases like correlation volumes and iterative warping with a single feed-forward encoder-decoder combining window, shifted-window, and reduced-resolution global attention. It reaches 0.68/1.48 EPE on Sintel Clean/Final, 3.23 Fl-all on KITTI-2015, and 3.192 1px on Spring. The architecture scales consistently from small to large variants and remains memory efficient at 1080p inference.
- Drops flow-specific components in favor of a pure transformer encoder-decoder
- Uses window, shifted-window, and global attention variants
- State-of-the-art on Sintel, KITTI-2015, and Spring benchmarks
- Memory-efficient at 1080p inference with consistent capacity scaling
Full article143 words · extracted from huggingface.co · click to collapse
Optical flow methods typically rely on task-specific inductive biases, such as correlation volumes, feature warping, and iterative refinement, among others, to reach high accuracy. While effective, such biases constrain the model to predefined heuristics, which can limit its expressivity and lead to more complex pipelines and additional computational cost. We present FreeFlow, a hierarchical transformer built without any flow-specific components, using instead a single feed-forward encoder--decoder. FreeFlow combines three attention variants: window attention for local processing, shifted-window attention for cross-window information exchange, and a global attention operating at a reduced resolution. The resulting architecture scales naturally with model capacity, enabling a consistent accuracy gain from small to large variants. Despite the absence of standard inductive biases, FreeFlow achieves state-of-the-art results on major benchmarks, including Sintel (0.68/1.48 EPE on Clean/Final), KITTI-2015 (3.23 Fl-all), and Spring (3.192 1px), while remaining memory efficient at 1080p inference.
Text extracted automatically; images, tables and formatting may be missing. Original: https://huggingface.co/papers/2609.11486