Technical details about generation
#1
by
tt84
- opened
Hi there, I wanted to know more about the technical details of how you generated these face images. Any information would be greatly appreciated. Did you use the OpenAI API or did you develop the dalle2 model yourself. If the latter, can you point me to any code?
From the paper, I think the DALL-E images were generated using the OpenAI API: https://arxiv.org/abs/2303.11408
For Stable Diffusion, I'm not sure but I imagine it's done with diffusers, using the v1.4 and v2 models from the Hub, like https://huggingface.co/stabilityai/stable-diffusion-2