What's the recommended Resize for RF Train Semantic Segmentation model?

I’m not sure what the recommended resize is for the Roboflow Semantic Segmentation model: You Might Be Resizing Your Images Incorrectly

We recommend a minimum Resize of 512x512 for the Roboflow Train Semantic Segmentation model. You can opt for a larger resize, but try to pick a size that is a multiple of 32 if you go higher than 512x512.

Another common choice is 1024x1024. Just note that with larger image sizes for training, you will also see longer training times, and slower inference.