We should warm up models, especially delegated to CoreML first, because the first run can take up to a few seconds.
We should warm up models, especially delegated to CoreML first, because the first run can take up to a few seconds.