In a blog post on Thursday, OpenAI announced the removal of the waitlist for its DALL E beta. This means that users can now register and use the image generator service immediately. OpenAI claims that more than 1.5 million users — artists, creative directors, authors and architects — actively create more than 2 million images a day using the tool. What’s more, more than 100,000 users share their feedback and creations with the company’s hostile community.
It is believed that responsibly scaling a system as powerful as DALL E, while exploring all creative uses and misuses, has required the tool to take an iterative approach. A deep learning image synthesis model, DALL E has been trained on millions of images ripped from the network. The latent diffusion method is used to learn associations between words and images. As a result, DALL E users enter a description called a ‘text cue’ and perceive it as a visually rendered image of 1024 1024 1024 pixels in any artistic style. In addition to uniquely generating the image into text, DALL E has introduced ‘Outpainting’, which allows users to upload an image by extending its boundaries with image synthesis. With this technique, multiple images can be combined into one, creating a visual bridge between them to blend styles.
OpenAI says, “We are currently testing the DALL E API with several customers and are excited to soon offer it more widely to developers and enterprises so they can build applications on this powerful system. We can’t wait to see what users around the world are creating with DALL E.”