Core-only variant of Lotus-2 Geometry. Skips the detail-sharpener
second stage entirely — the LCM output goes straight to the VAE decoder.
Mirrors Lotus-1's single-stage smoothness, on Lotus-2's improved base predictor.
No post-processing (no median de-grid) so the raw model behaviour is visible.
Returns previews + raw float32 .npy + 16-bit PNG for both depth and normal.
5122048
Examples
Image
Process resolution (detail cap). Output detail = min(input long side, this). Higher = finer but much more VRAM/time; >1536 may OOM on this GPU. Feed an input at least this large (upscale first) to benefit.