Instructions to use facebook/dinov3-vitl16-pretrain-lvd1689m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use facebook/dinov3-vitl16-pretrain-lvd1689m with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-feature-extraction", model="facebook/dinov3-vitl16-pretrain-lvd1689m")# Load model directly from transformers import AutoImageProcessor, AutoModel processor = AutoImageProcessor.from_pretrained("facebook/dinov3-vitl16-pretrain-lvd1689m") model = AutoModel.from_pretrained("facebook/dinov3-vitl16-pretrain-lvd1689m") - Notebooks
- Google Colab
- Kaggle
Request for access: Running TRELLIS.2-4B pipeline (Microsoft dependency)
Hello Meta/DINOv3 team,
My access request for facebook/dinov3-vitl16-pretrain-lvd1689m was
recently rejected. I would like to kindly ask for reconsideration.
Use case:
I am an independent software engineer working on a 3D asset generation
pipeline using Microsoft's TRELLIS.2-4B model. This DINOv3 ViT-L16
checkpoint is a hard dependency of TRELLIS.2 β it is referenced directly
in the pipeline config as the image conditioning encoder.
Why I need this specific checkpoint:
TRELLIS.2 was trained using dinov3-vitl16-pretrain-lvd1689m as the
image feature extractor. Replacing it with another DINOv3 variant or
DINOv2 would result in incompatible feature spaces and broken 3D output.
I agree to:
- Use the model solely for non-commercial research and development
- Not redistribute the model weights
- Comply with the model's license terms
Thank you for your consideration.