docs: remove PCA visualization guidance from Med3DINO model card

README.md

@@ -12,7 +12,7 @@ tags:
 
 **Three geometry-aware 3D medical encoders for representation learning and downstream research.**
 
-[Evaluation code](https://github.com/Endarzboy/REFORM3D) · [Project page](https://endarzboy.github.io/reform3d-website/) · [Checkpoint guide](https://github.com/Endarzboy/REFORM3D#checkpoints) · [PCA example](https://github.com/Endarzboy/REFORM3D/blob/main/notebooks/visualize_PCA.ipynb)
+[Evaluation code](https://github.com/Endarzboy/REFORM3D) · [Project page](https://endarzboy.github.io/reform3d-website/) · [Checkpoint guide](https://github.com/Endarzboy/REFORM3D#checkpoints)
 
 Med3DINO-ISO, REL and SA are self-supervised ViT-L/16 encoders trained from random initialization. The variants compare different geometric policies during pretraining.
 
@@ -80,8 +80,6 @@ For the registered `96³` evaluation inputs, prepare single-channel tensors with
 
 Spatial preprocessing follows the selected evaluation protocol, including its spacing, axis-order, cropping and resampling rules. Use the [configured evaluation pipeline](https://github.com/Endarzboy/REFORM3D/blob/main/docs/usage.md) for benchmark comparisons; resizing a scan to `96³` alone does not reproduce that preparation.
 
-For visualization, follow the [PCA notebook instructions](https://github.com/Endarzboy/REFORM3D#pca-visualization), including fetching the bundled CT/MR examples with Git LFS. All encoders share one CT crop or full-field MR input; MR visualization inputs can exceed `96³`. These viewing transforms are separate from benchmark preparation.
-
 ## Extract frozen features
 
 This example loads SA once and defines a function for preprocessed batches. The encoder is frozen and placed in evaluation mode by REFORM-3D.