Upload folder using huggingface_hub

UI: remove DA3 npz upload (not used)

UI: replace 3 separate file outputs with single Downloads component

fix: add /mv_sam3d /tmp /data to Gradio allowed_paths for output files

fix: download & symlink ALL checkpoint files (including .ckpt weights)

fix: force-add visualization source files ignored by MV-SAM3D .gitignore

fix: set LIDRA_SKIP_INIT=1 to bypass missing sam3d_objects/init.py

diag: add traceback to visualization __init__.py + strict build-time import test

fix: make utils3d install fail-loud (required dep), revert from my_requirements

fix: add missing utils3d dependency to my_requirements.txt

feat: two-stage accordion UI (Stage 1 multiview gen + Stage 2 3D recon)

Fix: add missing __init__.py in visualization/plotly + build-time import test

Move plotly/imageio/trimesh into my_requirements.txt

Fix: add plotly, imageio, trimesh missing from MV-SAM3D requirements

Fix: set CUDA_HOME and CONDA_PREFIX env vars for MV-SAM3D inference

Fix: install requirements line-by-line to avoid single-package abort

Add skip SAM-3 checkbox with white background removal option

Fix: strip trailing newline from FAL_KEY/HF_TOKEN env vars

Fix CMD: use env Python directly instead of micromamba run

Remove binary doc files, add to gitignore

update code

Update README with HF Space Docker SDK metadata

update code

Initial commit