|
Listen to this story
|
In a blog post on Thursday, OpenAI announced the removal of the waitlist for its DALL·E beta. This means that users can now immediately sign up and use the image generator service. OpenAI claims that more than 1.5 million users — artists, creative directors, authors, and architects — are actively creating over 2 million images a day with the tool. Moreover, over 100K users share their feedback and creations at the company’s discord community.
It is believed that responsibly scaling a system as powerful as DALL·E while learning about all the creative ways it can be used and misused has required the tool to have an iterative deployment approach. A deep learning image synthesis model, DALL·E has been trained on millions of images pulled from the web. A technique named latent diffusion is used to learn associations between the words and images. As a result, DALL·E users type in a description, called a ‘text prompt’ and see it as a visually rendered 1024×1024 pixel image in any artistic style of choice.
Besides the unique image-to-text generation, DALL·E introduced ‘Outpainting’, a feature which allows users to upload an image, extending its borders with the help of image synthesis. With this technique, several images can be merged into one by generating a visual bridge between them to blend styles.
OpenAI says, “We are currently testing a DALL·E API with several customers and are excited to soon offer it more broadly to developers and businesses so they can build apps on this powerful system. We can’t wait to see what users from around the world create with DALL·E.”
To know more about DALL·E beta, sign up here.
