DALL-E 2Translation site

What is DALL·E 2?

DALL-E 2 isOpenAILaunched in 2022, this image-generation model converts text descriptions into images and has also offered capabilities for local image editing and variant generation. It transforms subjects, compositions, and visual styles into image options that can be quickly compared, and has been used for concept design, illustration exploration, and creative sketching.

DALL·E 2 has been marked as deprecated by OpenAI and removed from the API.This status applies to the DALL·E 2 model and does not mean that all of OpenAI's image-generation services have been discontinued.

Three Types of Image Tasks

• Text Generation:Generate images based on descriptions of scenes, objects, and styles.
• Local Editing:Use a combination of images and text instructions to modify the specified area and explore adding or replacing visual elements.
• Image Variations:Generate different visual variations based on an existing image to compare composition and stylistic approaches.

Interface Migration

Legacy projects that still rely on this model need to be migrated to a supported image model, and the input formats, output specifications, editing capabilities, and costs should be reevaluated. Simply changing the model name does not guarantee the same results; output should be validated using the original business materials. You can start byOfficial Model DocumentationView the deprecation status and the current alternative entry point.