Show HN: Run TRELLIS.2 Image-to-3D generation natively on Apple Silicon

This is the kind of porting work I always hope for when I see a CUDA-only release. Have you thought about publishing the gather-scatter sparse 3D convolution and SDPA attention swaps as a standalone toolkit or writeup? A lot of folks running models locally on Apple Silicon hit the same wall with flash_attn, nvdiffrast, and custom sparse kernels and end up redoing the same work.

Most 'runs on Mac' ports are a wrapper around a cloud call or a quantized shell of the original model. Going after the CUDA-specific kernels with pure-PyTorch alternatives is the kind of work that ages well, because the next CUDA-locked research release is three weeks away. One question: how much of the gather-scatter sparse conv is reusable for other TRELLIS-like architectures, or is it bespoke to this one?

Cool project. I've been working on something similar in spirit — a personal video cloud (strimoza.com) — and the hardest part was also getting local playback to work reliably without internet. How are you handling memory pressure on the M-series chips with larger models?

I wonder how these outputs compare between Apple Sharp

https://github.com/apple/ml-sharp

No matter what it is cool seeing so much them work on different devices

Nice work. Although this model is not very good, I tried a lot of different image-to-3d models, the one from meshy.ai is the best, trellis is in the useless tier, really hope there could be some good open source models in this domain.

solid port. the sdpa swap for sparse attention — did you notice a meaningful quality difference, or is it basically equivalent to the cuda version? curious if the pure-pytorch path added any noticeable latency hit on the m3 max

This is fantastic, great work. I will attempt to run it on my 16GB M1 but I doubt it'll run.

Out of curiosity, how did you go about replacing the CUDA specific ops? Any resources you relied on or just experience? Would love to learn more.

So much effort, but no examples in the landing page.