Efficient Large-Scale Visual Representation Learning And Evaluation

Understand

Efficiently learning visual representations of items is vital for large-scale recommendations.

• In this article we compare several pretrained efficient backbone architectures, both in the convolutional neural network (CNN) and in the vision transformer (ViT) family.
• We describe challenges in e-commerce vision applications at scale and highlight methods to efficiently train, evaluate, and serve visual representations.
• We present ablation studies evaluating visual representations in several downstream tasks.