Update README.md

README.md

@@ -7,6 +7,8 @@ This encoder is a pure vision backbone designed for medical imaging foundation m
 
 ## Architecture Overview
 
+This encoder is initialized using `Qwen3-VL`'s vision encoder.  
+
 The implementation lives in `modeling_smb_vision.py` and exposes three main classes:
 
 - `SMBVisionEncoder`: Vision encoder with 3D patch embedding and stacked Transformer blocks
@@ -138,7 +140,10 @@ print(encoded_patches.shape)
 
 
 ## Training Data
-CTs for lung and abdomen from TCIA.  
+CTs for lung and abdomen from TCIA.
+TODO: Details about CTs.  
+TODO: Details about CXRs.  
+TODO: Details about MRIs.  
 
 
 ## Citation