To use Imagen on Vertex AI you must provide a text description of what you want to generate or edit. These descriptions are called prompts, and these prompts are the primary way you communicate with Generative AI on Vertex AI.
This guide shows you how modifying parts of a text-to-image prompt can produce different results and gives you examples of images you can create. This guide also provides guidance on how you can edit images using text prompts and iteration.
Product usage
To view usage standards and content restrictions associated with Imagen on Vertex AI, see the usage guidelines.
Content filtering - input text, uploaded images, and generated images
Generated images are filtered for undesirable or harmful content. Similarly, any input Imagen on Vertex AI receives is checked for offensive content. This includes the input text prompt and uploaded photos in the case of image editing. For more information, see Responsible AI and usage guidelines for Imagen.
You can also report suspected abuse of Imagen on Vertex AI or any generated output that contains inappropriate material or inaccurate information using the Report suspected abuse on Google Cloud form.
Prompt writing basics (subject, context, and style)
While there's no one way to write good prompts, adding some keywords and modifiers will help you get closer to your end goal. Prompts don't need to be long or complex, but most good prompts are descriptive and clear.
A good starting point can be to think of subject, context, and style.