- Link the benchmark datasets (ImageNet-1k-Depth, TartanGround, GraspNet) in the card metadata
- Usage: `pip install defm`, loading with `create_defm_model`, and a CNN example
- Model Zoo: corrected parameter counts and the ImageNet-1k-Depth numbers from the released benchmark code
- New Benchmarks section linking the benchmark code and datasets
18 days ago
Add root config.json (architecture of every variant) for download counting (#2)
- Update model card with paper link and sample usage snippets (207d01bfc369e87e21eed33e138fab6827b5cc1f)
Co-authored-by: Niels Rogge <nielsr@users.noreply.huggingface.co>
