Answer-first summary for fast verification
Answer: Use a complex model architecture, label with class labels, evaluate using precision and recall.
For a wide variety of dish presentations and cuisines, a complex model architecture is necessary to capture the intricate details. Labeling with class labels is appropriate for image classification tasks. Evaluating using precision and recall provides a balanced view of the model's performance, especially important in scenarios where false positives and false negatives have significant consequences.
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
No comments yet.
You are developing a custom image classification model for a food delivery platform that needs to classify images of dishes into different categories. The model needs to be highly accurate and capable of handling a wide variety of dish presentations and cuisines. Describe how you would approach the training of this model, including the choice of model architecture, the labeling process, and the metrics you would use to evaluate the model's performance.
A
Use a lightweight model architecture, label with object tags, evaluate using accuracy.
B
Use a complex model architecture, label with class labels, evaluate using precision and recall.
C
Use a lightweight model architecture, label with bounding boxes, evaluate using F1 score.
D
Use a complex model architecture, label with class labels, evaluate using mean average precision.