Update README.md

README.md

@@ -58,7 +58,6 @@ logits = model(**inputs).logits  # use top-k on tok.mask_token_id
 - **Layers / heads / width:** 19 encoder layers, 8 attention heads, hidden size 512; intermediate (MLP) size 768; GELU activations.  
 - **Attention:** Local window 128 with **global attention every 3 layers**; RoPE θ=160k (local & global).  
 - **Positional strategy:** `position_embedding_type: "sans_pos"`.  
-- **Dropout:** attention/embedding/MLP dropouts set to 0.0 in the published config.  
 
 ## Training data & procedure