Update README.md

README.md

@@ -63,7 +63,7 @@ This model is intended for research purposes in the field of neuropathology.
 ## Training Procedure
 
 * **Training System/Framework:** DINO-MX (Modular & Flexible Self-Supervised Training Framework)
-* **Training Strategy:** Hieararchically connected tiles
+* **Training Strategy:** The model is trained using *Magnification-Aware Distillation (MAD)*, where a teacher network processes low-magnification tiles (2.5× or 10×) to capture global tissue context, while a student network learns from spatially aligned high-magnification tiles (10× or 40×) to encode fine-grained cellular details. Training is performed in a pairwise manner (2.5×→10× and 10×→40×), enforcing semantic consistency across resolutions and enabling unified cross-magnification representation learning.
 * **Training Infrastructure:** 4 x DGS H100 nodes (32 x H100 GPUs) 
 * **Base Model (if fine-tuning):** Pretrained `facebook/dinov2-with-registers-giant` loaded from Hugging Face Hub.
 * **Training Objective(s):** Self-supervised learning using DINO loss, iBOT masked-image modeling loss.