OpenAI's DALL-E service has received several updates that will help make changes to generated pictures. In the paid subscription of ChatGPT, there is a built-in tool for drawings, and now when using DALL-E, users can selectively edit the generated output.

To do this, you need to create an illustration in a separate function, then click on the result, and a panel with tools will appear. Click on the palette, select the area that needs to be corrected (you don't need to be precise with the contour, the neural network will understand). Then, in the chat window on the right, write what should be in the new version. For example, change a plate of rice to fried chicken wings. The new option will be useful for iterative design, personalization, neural network training, and collaborative work. It will simplify the selection of options and ideas for the final visual.

Dall-E 3Image: EnsiplayRecently, hints have appeared in the drawing tool for ChatGPT 4. You can select the orientation of the graphic by clicking buttons: square, horizontal, or vertical. There are also styles like anime, fantasy, and something unusual like Gold. This is useful for quickly choosing the desired format and style without having to waste space in the text prompt and manually specifying everything. It will save time and simplify the generation process.

Main image: Ensiplay

Comment on this