TRELLIS.2 Text & Image to 3D in the Browser with SpawnDev.ILGPU.ML (v5.3.12)

#68
by LostBeard - opened

Hey everyone,

I wanted to share that SpawnDev.ILGPU.ML now supports full Text & Image to 3D generation using Microsoft's TRELLIS.2 (4B), running 100% client-side in the browser via WebGPU!

Everything from SD Turbo to RMBG-1.4, the DINOv3 feature extractors, and the three 1.3B-parameter flow transformers executes entirely on your own GPU via C# and WebGPU, with zero server uploads required.

Check out the release and demo:

Let me know what you think!
2026-10-17-16-55_TextTo3D-Chest

To take it a step further, we also chained SkinTokens (for neural auto-rigging and learned skin weights) and Kimodo SOMA (for skeletal motion retargeting and animation) into the same client-side browser pipeline.
2026-10-19_TextTo3D-Rig-Animate-Knight-FLUX2-1024

طبق كشري

طبق كشري مصري

pngtree-egyptian-koshari-bowl-with-rice-lentils-pasta-crispy-onions-and-tomato-png-image_16176599

Sign up or log in to comment