Update app.py

app.py

@@ -123,8 +123,12 @@ def _infer_image(image_path: str) -> str:
         torch.cuda.empty_cache()
 
         with torch.no_grad():
-            ids = model.generate(**inputs, max_new_tokens=MAX_NEW_TOKENS)
-
+            ids = model.generate(
+                **inputs,
+                max_new_tokens=MAX_NEW_TOKENS,
+                do_sample=False,
+                repetition_penalty=1.1,
+            )
         result = processor.decode(
             ids[0][inputs["input_ids"].shape[1]:],
             skip_special_tokens=True,