A mordernized ViT-Adapter under maintenance. No old and heavy libraries, just the latest pytorch and mordern packages.
computer-vision cuda object-detection vit lightweight-framework semantic-segmentation instance-segmentation pytorch-implementation optimized-code mmdetection vision-transformer modernized dinov3 vit-adapter maskdino
-
Updated
May 6, 2026 - Python