Module 5: Using pre-trained models and embeddings

Using pre-trained models and embeddings

Slides

View slides in full screen

Outline

• Pre-trained models and embeddings
• Image emebddings: Image classification and clustering with pretrained CNNs (e.g., DenseNetNet)
• Text embeddings: Document clustering with MiniLM
• Brief demo or mention of object detection (YOLO)
• CLIP: text-to-image and image-to-text similarity