Image Feature Extraction
timm
Safetensors
English
pathology
histopathology
feature-extraction
spectra
Instructions to use medarc/OpenMidnight-1.1-SPECTRA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- timm
How to use medarc/OpenMidnight-1.1-SPECTRA with timm:
import timm model = timm.create_model("hf-hub:medarc/OpenMidnight-1.1-SPECTRA", pretrained=True) - Notebooks
- Google Colab
- Kaggle
Commit History
3 days ago
-
Use full-image resize for all input shapes
Verified signatureSet crop_mode=squash alongside crop_pct=1.0 so the standard timm transform directly resizes the entire image to 224x224. Simplify the model-card instructions. Square-tile outputs are unchanged; verified bit-for-bit equivalence to explicit resizing for square and rectangular images with timm 1.0.30.ac9652c -
Preserve full square tiles in preprocessing
Verified signatureSet crop_pct to 1.0 so 256x256 tiles resize to 224x224 with the full field of view. Document segmentation mask alignment and the need to recompute embeddings and retrain decoders after this correction. Validated the transform against direct full-tile resizing with timm 1.0.30.2fa7a84
22 days ago
-
Publish selected full checkpoint and concise inference model card
Verified signature0fe82c7