Watch this video on YouTube
After deploying several versions of an image classification model on AI Platform, you aim to track and compare their performance over time. What approach should you take to effectively perform this comparison?