Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload requirements.txt with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload requirements.txt with huggingface_hub

Upload licenses/LICENSE-RMBG2-CC-BY-NC-4.0 with huggingface_hub

Upload licenses/LICENSE-BiRefNet-MIT with huggingface_hub

Upload licenses/LICENSE-DINOV2-APACHE2 with huggingface_hub

Upload licenses/LICENSE-TRELLIS2-MIT with huggingface_hub

Upload MODEL_LICENSES.md with huggingface_hub

Upload README.md with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Add histogram normalization on foreground pixels after alpha premultiply

Best-of-N noise selection: probe 3 candidates, pick lowest std

Revert to premultiplied alpha on black (original preprocessing)

Fix darkened texture: composite on white background instead of black

Restore premultiplied alpha preprocessing (matches training data)

Move Generate and Extract GLB buttons above preview/GLB windows

Fix darkened texture: lower alpha threshold from 0.5 to 0.05

Fix GLB export: remove hidden visibility, return plain paths

Restore logo layout + fix GLB export with DownloadButton

Fix degenerate detection: bypass multi-image patching during probe

Fix GLB export: add DownloadButton and return tuple

Fix black texture bug: detect degenerate sampler trajectories and retry

Suppress third-party warnings and add render diagnostics

Fix dark texture: revert tex_mode default to stochastic and restore alpha preprocessing

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Fix o_voxel import: monkey-patch postprocess instead of sys.path override

Add separate texture algorithm selector (default: multidiffusion)

Fix multidiffusion CFG and add per-stage texture mode

Add sys.path override to use local o-voxel with texture fix