Instructions to use xocialize/trellis2-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use xocialize/trellis2-mlx with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir trellis2-mlx xocialize/trellis2-mlx
- Trellis
How to use xocialize/trellis2-mlx with Trellis:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
You need to agree to share your contact information to access this model
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
This repository redistributes the DINOv3 image conditioner (dino.safetensors) under the DINOv3 License (see DINOv3_LICENSE.md). To access these weights you must agree to the DINOv3 License, including its Acceptable Use restrictions and Trade Control terms. The remaining components are Microsoft TRELLIS.2 (MIT).
Log in or Sign Up to review the conditions and access this model content.
TRELLIS.2 β MLX (consolidated)
MLX-ready, consolidated weights for Pixal3D / Microsoft TRELLIS.2 single-image β textured 3D mesh,
for the Swift-MLX port xocialize/mlx-trellis2-swift
(MLXEngine imageTo3D). One gated download replaces three upstream repos, and the keys are already
in the port's module layout, so the Swift loader consumes them remap-free.
β οΈ Gated β includes DINOv3.
dino.safetensorsis the DINOv3 conditioner (Meta), redistributed here under the DINOv3 License (Β§1.b permits redistribution under its terms; this repo is gated so downloaders accept the license, and ships a copy of it). By accessing this repo you agree to the DINOv3 License β including its Acceptable Use and Trade Control terms. The TRELLIS.2 components are MIT (LICENSE).
Contents
| File | Component | License | dtype |
|---|---|---|---|
dino.safetensors |
DINOv3 ViT-L/16 image conditioner | DINOv3 License | bf16 |
struct_flow.safetensors |
sparse-structure flow DiT | MIT | bf16 |
struct_dec.safetensors |
sparse-structure decoder | MIT | fp16 |
shape_flow_512.safetensors / shape_flow_1024.safetensors |
shape SLat flow DiT (512 / 1024 tiers) | MIT | bf16 |
shape_dec.safetensors |
shape SLat decoder (FlexiDualGrid) | MIT | fp16 |
tex_flow_512.safetensors / tex_flow_1024.safetensors |
texture SLat flow DiT (512 / 1024 tiers) | MIT | bf16 |
tex_dec.safetensors |
texture SLat decoder (PBR) | MIT | fp16 |
normalization.json |
shape + texture SLat mean/std | β | β |
Total β 15 GB. The _1024 flows are only needed for the res1024 / res1536 cascade tiers. Tensor
values are byte-identical to upstream; only keys/layout were adapted to the Swift module namespace.
Usage
hf download xocialize/trellis2-mlx --local-dir trellis2-mlx # accept the DINOv3 License when prompted
The Swift port loads every file directly (no key remap). Geometry + per-vertex texture color;
res512 / res1024 / res1536 resolution tiers.
Provenance & license
Derived from microsoft/TRELLIS.2-4B + microsoft/TRELLIS-image-large (MIT, see LICENSE) and
facebook/dinov3-vitl16-pretrain-lvd1689m (DINOv3 License, see NOTICE).
The most-restrictive component (DINOv3) governs access β hence the gate. Port code is MIT.
Quantized