Today I want to share a video I recorded of an experiment I conducted using ChatGPT to create SVG images. As many of you may know, GPT-3 has shown great potential for creating text, but I was curious to see if it could also create images.
To test this, I used ChatGPT, which is a variant of GPT-3 that is optimized for conversation and responding to natural language inputs. I gave ChatGPT a prompt to create an SVG image, and the results were quite impressive.
The images generated by ChatGPT were a clear improvement compared to the ones created by text-davinci-003, and even showed progress from text-davinci-002. However, they still aren’t quite good enough for practical use.
Despite this, I think the potential for using large language models like GPT-3 for image generation is exciting, and I can’t wait to see what the future holds in this field.
Check out the video below to see ChatGPT in action, and let me know what you think in the comments.
Thanks for reading, and be sure to check back for more updates on my experiments with GPT-3 and image generation.