ProductIntelHub
Sign In

TRELLIS.2 image-to-3D running on Mac Silicon – no Nvidia GPU needed

  • Hacker News

I ported Microsoft's TRELLIS.2 (4B parameter image-to-3D model) to run on Apple Silicon via PyTorch MPS. The original requires CUDA with flash_attn, nvdiffrast, and custom sparse convolution kernels: none of which work on Mac. I replaced the CUDA-specific ops with pure-PyTorch...

  • Published: Apr 20, 2026
  • First seen: Apr 20, 2026
Visit website

AI Summary

I ported Microsoft's TRELLIS.2 (4B parameter image-to-3D model) to run on Apple Silicon via PyTorch MPS. The original requires CUDA with flash_attn, nvdiffrast, and custom sparse convolution kernels: none of which work on Mac. I replaced the CUDA-specific ops with pure-PyTorch...

Best for

Teams evaluating AI product workflows / Builders comparing emerging tools / Operators tracking early category shifts

Why it matters

Primary discovery source is Hacker News.

Key Features

  • Primary public product URL is https://github.com/shivampkumar/trellis-mac.
  • Description: I ported Microsoft's TRELLIS.2 (4B parameter image-to-3D model) to run on Apple Silicon via PyTorch MPS. The original requires CUDA with flash_attn, nvdiffrast, and custom sparse convolution kernels: none of which wor....
  • GitHub repository is linked as shivampkumar/trellis-mac.
  • Listed on Hacker News as "TRELLIS.2 image-to-3D running on Mac Silicon – no Nvidia GPU needed".
  • Source description: I ported Microsoft's TRELLIS.2 (4B parameter image-to-3D model) to run on Apple Silicon via PyTorch MPS. The original requires CUDA with flash_attn, nvdiffrast, and custom sparse convolution kernels: none of which wor....

Use Cases

  • Primary discovery source is Hacker News.
  • A public GitHub repo is available for direct technical review.
  • Hacker News mention is recent (2026-04-20).
  • Primary public product URL is https://github.com/shivampkumar/trellis-mac.
  • Description: I ported Microsoft's TRELLIS.2 (4B parameter image-to-3D model) to run on Apple Silicon via PyTorch MPS. The original requires CUDA with flash_attn, nvdiffrast, and custom sparse convolution kernels: none of which wor....

Why Now

TRELLIS.2 image-to-3D running on Mac Silicon – no Nvidia GPU needed is appearing on fresh discovery surfaces, so it is worth reviewing while momentum is still forming. Confidence is currently medium (49/100), so treat this as an early signal rather than a settled trend.

Community Signals

Trend score

34.3

24h momentum

Rising

Hacker News points

65

Rising

Facts / Signals / Inference / Unknowns

Facts

  • Listed on Hacker News as "TRELLIS.2 image-to-3D running on Mac Silicon – no Nvidia GPU needed".
  • Source description: I ported Microsoft's TRELLIS.2 (4B parameter image-to-3D model) to run on Apple Silicon via PyTorch MPS. The original requires CUDA with flash_attn, nvdiffrast, and custom sparse convolution kernels: none of which wor....
  • Source publish date is 2026-04-20.
  • Description: I ported Microsoft's TRELLIS.2 (4B parameter image-to-3D model) to run on Apple Silicon via PyTorch MPS. The original requires CUDA with flash_attn, nvdiffrast, and custom sparse convolution kernels: none of which wor....
  • GitHub repository is linked as shivampkumar/trellis-mac.
  • Primary public product URL is https://github.com/shivampkumar/trellis-mac.

Signals

  • Hacker News mention is recent (2026-04-20).
  • A public GitHub repo is available for direct technical review.
  • Primary discovery source is Hacker News.

Inference

  • Public code access can lower evaluation friction for developer audiences.

Unknowns

  • Documentation is not explicitly linked in the current allowed evidence set.
  • No tagline is stored on the current product record.
  • Pricing details are not explicitly linked in the current allowed evidence set.
  • Recent changelog or release history is not explicitly linked in the current allowed evidence set.
  • Release cadence cannot be confirmed unless a changelog or release link is explicitly provided.

Evidence Snapshots

TRELLIS.2 image-to-3D running on Mac Silicon – no Nvidia GPU needed

Listed on Hacker News as "TRELLIS.2 image-to-3D running on Mac Silicon – no Nvidia GPU needed".

TRELLIS.2 image-to-3D running on Mac Silicon – no Nvidia GPU needed GitHub repository

GitHub repository is linked as shivampkumar/trellis-mac.

TRELLIS.2 image-to-3D running on Mac Silicon – no Nvidia GPU needed official profile

Primary public product URL is https://github.com/shivampkumar/trellis-mac.

Alternatives / Related

Sfsym – Export Apple SF Symbols as Vector SVG/PDF/PNG

Hacker News

I found myself reaching for SF Symbols' 'Copy Image As…' quite often during agentic design sessions, so I made a command-line tool that the agent can use by itself. It exports Apple SF Symbols as SVG, PDF, or PNG. The vector paths come directly from macOS's symbol renderer. In...

Open detail

Run GUIs as Scripts

Hacker News

Hi there, Zero Stars here. I recently published some new work to Hokusai Pocket, which is a cross-platform binary made on top of raylib and MRuby that runs GUIs from ruby scripts. License? MIT! How does it work? The binary is available on the GitHub releases page: https://gith...

Open detail

A stateful UI runtime for reactive web apps in Go

Hacker News

Doors: Server-driven UI framework + runtime for building stateful, reactive web applications in Go. Some highlights: * Front-end framework capabilities in server-side Go. Reactive state primitives, dynamic routing, composable components. * No public API layer. No endpoint desi...

Open detail

Continual Learning with .md

Hacker News

I have a proposal that addresses long-term memory problems for LLMs when new data arrives continuously (cheaply!). The program involves no code, but two Markdown files. For retrieval, there is a semantic filesystem that makes it easy for LLMs to search using shell commands. It...

Open detail

Original Sources

Hacker News