Update GitHub URL to UVA-Computer-Vision-Lab/sab-3r

Keep upsampler on CPU on HF Spaces (fixes text-query _cuda_init crash) — verified locally

ZeroGPU: patch cached device descriptors (working_device) to CPU after roundtrip — verified locally

ZeroGPU: force return value to CPU via torch.save/load roundtrip

ZeroGPU: move tensors to CPU on return; drop @GPU from post-processing; harden SparseGAState.__del__

Move SAB3R model to module-level singleton for ZeroGPU (fixes CPU/CUDA weight mismatch on worker)

Add ZeroGPU support (@spaces.GPU decorators)

Make demo device-adaptive (cuda if available, else cpu)

Initial SAB3R demo release