TRELLIS.2
TRELLIS.2 is Microsoft's open 4B image-to-3D model. It returns the object's shape as a mesh, on a grid of up to 1536^3. LibreYOLO runs the same weights in plain PyTorch through the LibreGen factory, geometry only.
- Задачи
- 3d generation
- Размеры
- 4b at 1024 px
- Установка
pip install "libreyolo[gen]"- Уровень поддержки
- Смежный уровень, начиная с v1.8.0. Отдельная часть продукта со своей фабрикой и контрактом.
- Исходный проект
- TRELLIS.2 от Microsoft Research, DINOv3 License. Статья, исходный код
- Лицензии
- Код: MIT, веса: MIT and DINOv3 License. Коммерческое использование
Install
pip install "libreyolo[gen]"The extra brings transformers, which loads the DINOv3 image encoder.
Predict
Nothing is gated and no login is needed. Weights download on first use at pinned revisions.
from libreyolo import LibreGen, SAMPLE_IMAGE model = LibreGen("trellis2") # also "trellis.2"result = model(SAMPLE_IMAGE, seed=0) result.save("object.obj")m = result.meshesprint(m.vertices[0].shape, m.faces[0].shape, m.seed)# Saves the object as .obj under runs/generate/predict.libreyolo generate source=https://raw.githubusercontent.com/LibreYOLO/libreyolo/release/libreyolo/assets/guggenheim-bilbao.jpg model=trellis2 seed=0from libreyolo import LibreGen, SAMPLE_IMAGE model = LibreGen("trellis2")result = model(SAMPLE_IMAGE, resolution=512) # one shape stage, fasterresult = model(SAMPLE_IMAGE, resolution=1536) # finest gridA structure stage places the occupied voxels, shape stages fill in a latent on them (at 512, then again at 1024 or 1536), and a sparse decoder turns it into a dual-grid mesh. Both flow stages read DINOv3 ViT-L/16 tokens. Upstream needs CUDA-only extensions (FlexGEMM, flash-attn, o-voxel); LibreYOLO runs the same weights in plain PyTorch on NVIDIA GPUs, Apple silicon and CPU. Noise is drawn on the CPU, so a seed gives the same object on every device up to kernel rounding.
result.meshes holds the mesh: vertices (V, 3), z-up in [-0.5, 0.5]^3,
and faces (F, 3). It is geometry only and the raw dual grid, before
upstream's CuMesh clean-up: faces are not consistently wound and some vertices
are unused. Upstream's texture stage bakes with nvdiffrast, a non-commercial
NVIDIA library, and is not part of LibreYOLO.
On an A100 the library measured 19.5 s and 14.3 GB peak GPU memory for one object at 1024, and 5.5 s and 11.6 GB at 512.
Settings
| Setting | Default | Meaning |
|---|---|---|
steps | 12 | Sampling steps. |
guidance | 7.5 | Classifier-free guidance. |
resolution | 1024 | Output grid: 512, 1024 or 1536. |
margin | 1.0 | Crop side over object size. |
max_tokens | 49152 | Token budget for the cascade; it may lower the resolution to fit, and params records the resolution used. |
Variants
One size, 4b: 4B parameters in five networks, about 8.7 GB, plus 1.2 GB for
the DINOv3 encoder and 1.3 GB from the TRELLIS mirror.
Export
| Задача | ONNX | TorchScript | ExecuTorch | TensorRT | OpenVINO | Paddle | MNN | RKNN | Axelera | ncnn | TFLite | CoreML | Core AI |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 3D generation | 3D generation to ONNX: не поддерживается | 3D generation to TorchScript: не поддерживается | 3D generation to ExecuTorch: не поддерживается | 3D generation to TensorRT: не поддерживается | 3D generation to OpenVINO: не поддерживается | 3D generation to Paddle: не поддерживается | 3D generation to MNN: не поддерживается | 3D generation to RKNN: не поддерживается | 3D generation to Axelera: не поддерживается | 3D generation to ncnn: не поддерживается | 3D generation to TFLite: не поддерживается | 3D generation to CoreML: не поддерживается | 3D generation to Core AI: не поддерживается |
Export is not implemented for LibreGen 3D families. export(), train() and
val() raise NotImplementedError (ADR 0031 in the library).
Licensing
Проверяйте лицензию в репозитории конкретных весов на Hugging Face. Она указана для каждого чекпойнта в организации LibreYOLO и может различаться даже внутри одного семейства. Этот репозиторий считается авторитетным источником, а сводка ниже описывает условия на момент последней проверки страницы.
Это описание соответствующих лицензий, а не юридическая консультация. Если ответ важен для коммерческого использования, прочитайте лицензии самостоятельно и проконсультируйтесь с юристом.
- Оригинальная работа
- TRELLIS.2, Microsoft Research
- Лицензия исходного проекта
- DINOv3 License
- Исходный код проекта
- github.com/microsoft/TRELLIS.2
- Код LibreYOLO
- MIT
- Веса
- MIT and DINOv3 License, повторно опубликованы на huggingface.co/LibreYOLO
- Толкование
- The TRELLIS.2 code and its microsoft/TRELLIS.2-4B weights are MIT, a permissive license that allows commercial and closed-source use with the copyright notice kept. The image encoder it needs is not: facebook/dinov3-vitl16-pretrain-lvd1689m is Meta's, under the DINOv3 License, a custom agreement that is not OSI-approved. It permits redistribution with a copy of the agreement and passes through trade-control and end-use restrictions (military, weapons, espionage and nuclear uses among them), so it sets the terms for the pipeline as a whole; read it before shipping a product on it. LibreYOLO's adapter is MIT, and the flow models, VAE decoder and dual-grid meshing it adapts from TRELLIS.2 keep MIT, noted in libreyolo/models/gen/trellis2/NOTICE. Upstream's nvdiffrast and nvdiffrec (NVIDIA, non-commercial) and BRIA's RMBG-2.0 (non-commercial) are not used, and the texture stage is not ported. Weights are redistributed unchanged at pinned revisions and none is gated: the geometry networks as LibreYOLO/LibreTRELLIS2-4b, TRELLIS's structure decoder and the BiRefNet weights (MIT) from LibreYOLO/LibreTRELLISl, and the DINOv3 encoder as LibreYOLO/LibreDINOv3vitl16 with the agreement.
TRELLIS.2 is by Microsoft Research
(microsoft/TRELLIS.2). The code and
the microsoft/TRELLIS.2-4B weights are MIT.
The image encoder sets the terms for the pipeline. It is Meta's
facebook/dinov3-vitl16-pretrain-lvd1689m, under the DINOv3 License, a
custom agreement that is not OSI-approved and passes through trade-control and
end-use restrictions (military, weapons, espionage and nuclear uses among
them). Read it before shipping a product on this model.
Weights are redistributed unchanged at pinned revisions and none is gated: the geometry networks as LibreYOLO/LibreTRELLIS2-4b, TRELLIS's structure decoder and BiRefNet (MIT) from LibreYOLO/LibreTRELLISl, and the DINOv3 encoder as LibreYOLO/LibreDINOv3vitl16 with the agreement. nvdiffrast, nvdiffrec and RMBG-2.0 (all non-commercial) are not used.
Citation
@article{
xiang2025trellis2,
title={Native and Compact Structured Latents for 3D Generation},
author={Xiang, Jianfeng and Chen, Xiaoxue and Xu, Sicheng and Wang, Ruicheng and Lv, Zelong and Deng, Yu and Zhu, Hongyuan and Dong, Yue and Zhao, Hao and Yuan, Nicholas Jing and Yang, Jiaolong},
journal={Tech report},
year={2025}
}Скопировано из блока цитирования авторов на странице github.com/microsoft/TRELLIS.2#-citation.