change max_new_tokens whisper

app.py

@@ -72,7 +72,7 @@ pipe = pipeline(
     batch_size=64,  # batch size for inference - set based on your device
     torch_dtype=torch_dtype,
     device=device,
-    max_new_tokens=512,  # Limit text generation
+    max_new_tokens=444,  # 448 max - 4 prompt tokens = 444
     return_timestamps=False,  # Save memory
 )