Is the issue with faces a deliberate choice by the devs?
I recreated famous album covers with DALL-E
131–137 of 137 posts
Re: I recreated famous album covers with DALL-E
#132Man, after seeing Stable Diffusion's output, DALL-E's looks just janky. Like watching a propeller plane after seeing a jet. Crazy how fast the tech is moving.
DALL-E is capable of very high quality photorealistic images with the right prompts. Here's one I made: https://imgur.com/yAzKkHb “High detail, macro portrait photo, a handsome Australian man with a strong jaw line, blue eyes and brown hair, smiles at the camera, set in an outdoor pub at golden hour, shot using a ZEISS Supreme Prime lens”
https://reddit.com/r/dalle2/comments/wsi97q/_/ikyjqhh/?conte...
> High detail, macro portrait photo, a [physical descriptor, regional identity, etc] man/woman with [eye color] eyes and [hair color] hair, smiles at the camera, set in a [field/dimly lit room/whatever] at golden hour, shot using a ZEISS Supreme Prime lens
The suggestions in that thread are quite effective, particularly the notion of reducing synthetic ‘beauty’ for a more human appearance.
Re: I recreated famous album covers with DALL-E
#133Earlier quoted context omitted.
How long until an unrestricted DALL-E -level model emerges? Seems only a matter of time.
Stable Diffusion is what you're looking for. It should be released to the public quite soon actually.
Re: I recreated famous album covers with DALL-E
#134Re: I recreated famous album covers with DALL-E
#135I wonder, how much energy is being burned for these kinds of experiments.
Re: I recreated famous album covers with DALL-E
#136Earlier quoted context omitted.
Stable Diffusion is what you're looking for. It should be released to the public quite soon actually.
Yep, here’s interview with the founder of Stable AI, which includes their intent to release Stable Diffusion and be more open than OpenAI: https://m.youtube.com/watch?v=YQ2QtKcK2dA