Update src/model_loader.py

src/model_loader.py

@@ -1,14 +1,4 @@
 # src/model_loader.py
-# -----------------------------------------------------------------------------
-# Why this change
-# -----------------------------------------------------------------------------
-# - Fix fallback model id → 'google/medgemma-4b-text-it' (previous typo caused
-#   CPU-only runs to fail).
-# - Keep primary on GPU in 4-bit (bnb, nf4) when available; otherwise fallback.
-# - Provide a single generate_chat(messages, **gen_kwargs) entry point with
-#   consistent logging and without relying on chat templates (manual prompt).
-# - Lightweight logs show model choice, cache path, and generation time.
-# -----------------------------------------------------------------------------
 
 from __future__ import annotations