Enroll now: bit.ly/49YBjLb
Today we’re releasing Prompt Engineering for Vision Models, a short course built in collaboration with Comet.
Prompt engineering is also used in vision models, and depending on the model, prompts can be text, pixel coordinates, bounding boxes, or segmentation masks.
Learn to prompt vision models like Meta's Segment Anything Model (SAM), a universal image segmentation model, OWL-ViT, a zero-shot object detection model, and Stable Diffusion 2.0, a widely used diffusion model, and use a fine-tuning technique called DreamBooth to tune a diffusion model to associate a text label with an object of your preference.
You’ll explore:
- Image Generation: Prompt with text and by adjusting hyperparameters like strength, guidance scale, and number of inference steps.
- Image Segmentation: Prompt with positive, negative, and bounding box coordinates.
- Object detection: Prompt with text to produce a bounding box to isolate specific objects within images.
- In-painting: Combine the above techniques to replace objects within an image with generated content.
- Personalization with Fine-tuning: Use DreamBooth, a fine-tuning technique to generate custom images based on pictures that you provide.
- Iterating and Experiment Tracking: Prompting and hyperparameter tuning are iterative processes, and experiment tracking can help to identify effective combinations. This course will use Comet, a library to track experiments and optimize visual prompt engineering workflows.
Get started with prompting vision models today.
Learn more: bit.ly/49YBjLb
28 апр 2024