fix(dino): drop default HF near-assets; use torch.hub facebookresearch/dinov2

feat(app): block Gradio bind until CPU weights load (no background preload)

fix(spaces): Hdri torch.load after ConvNeXt; default off CPU preload with spaces

fix(load): CPU-only checkpoint load for ZeroGPU main process

fix(pipeline): do not treat ckpts/... as Hugging Face org/model id

app: CPU preload Hunyuan+NeAR, cuda move in GPU callbacks; drop gsplat warmup

feat: add app_gsplat ZeroGPU probe (RGB+ED raster only)

refactor(hyshape): CPU preload at startup, GPU move on Generate Mesh

feat(hyshape): GPU geometry warmup on page load (ZeroGPU cold start)

feat(space): HyShape-only ZeroGPU diagnostic app as Space entry

fix(space): split ZeroGPU cold-start paths and add timing diagnostics