Question about inference_res in inference.py

I've noticed that in the **inference.py** file, the variable **inference_res** seems to be independent of the input video's resolution. I'm wondering if I can modify it without affecting the model's performance. The videos I need to process have a resolution of only 224*224.