The objective at the model iteration stage is to train a model to achieve the best possible performance learning from our annotated dataset. The ultimate goal is a model reaching human-level performance. Such a model can be applied to visual recognition tasks such as self-driving cars.
Model iteration trains models with the annotated dataset and then evaluates the model’s performance. If a model performs great, we can deploy it into production. If the performance is poor, we analyze the root causes to find ways to improve it.
Based on our past experience at Landing AI we have developed best practices for model training and evaluation. In the following article, I share a few high-priority tasks during model training. We openly share our guiding principles to help machine learning engineers (MLEs) through model training and evaluation.