Instructions to use microsoft/colipri with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- COLIPRI
How to use microsoft/colipri with COLIPRI:
pip install colipri
from colipri import get_model from colipri import get_processor from colipri import load_sample_ct from colipri import ZeroShotImageClassificationPipeline model = get_model().cuda() processor = get_processor() pipeline = ZeroShotImageClassificationPipeline(model, processor) image = load_sample_ct() pipeline(image, ["No lung nodules", "Lung nodules"])
- Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -17,9 +17,9 @@ COLIPRI is a 3D vision–language transformer model trained to encode chest
|
|
| 17 |
<!-- Provide a longer summary of what this model is. -->
|
| 18 |
|
| 19 |
COLIPRI was trained using tens of thousands of chest CT scans and reports, without any annotations, using multiple objectives to learn strong joint representations of 3D images and text.
|
| 20 |
-
The procedure is described in detail in
|
| 21 |
|
| 22 |
-
The weights shared here correspond to our best-performing model
|
| 23 |
|
| 24 |
- **Developed by:** Microsoft Research
|
| 25 |
- **Model type:** 3D vision–language encoder
|
|
|
|
| 17 |
<!-- Provide a longer summary of what this model is. -->
|
| 18 |
|
| 19 |
COLIPRI was trained using tens of thousands of chest CT scans and reports, without any annotations, using multiple objectives to learn strong joint representations of 3D images and text.
|
| 20 |
+
The procedure is described in detail in [Comprehensive language-image pre-training for 3D medical image understanding](https://link.springer.com/chapter/10.1007/978-3-032-37624-4_24) (Wald et al., ECCV 2026).
|
| 21 |
|
| 22 |
+
The weights shared here correspond to COLIPRI-CRM, our best-performing model in the paper.
|
| 23 |
|
| 24 |
- **Developed by:** Microsoft Research
|
| 25 |
- **Model type:** 3D vision–language encoder
|