Fix: remove unsupported max_model_len param for PyPI package

main.py

@@ -326,10 +326,9 @@ async def startup():
     print("Loading Orpheus model (this may take a moment)...")
     from orpheus_tts import OrpheusModel
     
-    model = OrpheusModel(
-        model_name=ORPHEUS_MODEL,
-        max_model_len=MAX_MODEL_LEN
-    )
+    # Note: PyPI orpheus-speech 0.1.0 uses simpler API
+    # model_name can be "medium-3b" or full HF path
+    model = OrpheusModel(model_name=ORPHEUS_MODEL)
     
     print("✓ Orpheus model loaded successfully")