Spaces:
Running on Zero
Running on Zero
Commit History
-
fix: match v5 checkpoint architecture and benchmark scaled inference
Verified signature5878871 -
Expose adaptive-bin uncertainty outputs
Verified signatureDerive per-pixel height uncertainty from the adaptive-bin distribution, propagate it through tiled inference and TTA, and publish ndsm_std_m.npy for the DepthWizard frontend.98ddc8c
-
Enhance GPU processing and batching capabilities; update README for clarity on test-time augmentation and batch sizes
17aade5
-
Add DINOv3 model components and loss functions
- Implemented DPT trunk architecture in models/dpt.py for feature extraction. - Added exponential moving average functionality in models/ema.py to stabilize training. - Created DINOv3 encoder in models/encoder.py for robust block finding and layer-wise learning rate decay. - Developed heads for regression and classification tasks in models/heads.py, including gated fusion mechanism. - Introduced custom loss functions in models/losses.py to address specific training challenges. - Implemented test-time augmentation strategies in models/tta.py to enhance model predictions. - Updated requirements.txt to include necessary dependencies for the new model components.40d7c69