High Resolution Text to Image with Retrieval Augmented Diffusion!

Retrieval Augmented Diffusion Models are the latest thing in the Latent Diffusion repository, just while we wait for a public stable diffusion model release! At 768x768 and using ViT-L/14 the results can be pretty amazing!

Also included is a large art style study, so if you want to know which is your favourite style then be sure to check it out 😁

Links:

Chapters:
00:00 - Intro
00:10 - Basic install & usage
07:50 - Databases & scales
15:25 - Art style comparison

[tape-titans]

You told me I didn't need a 3090! Just wait for the 4000 series he said! And now I have to look at all this cool stuff. Mean.

[YandereShiki]

That's alot of codes to make realistic oil paintings.You got extreme skills.

[debtwolf]

You sound like David Brent from the Office :) I like it!

[petecircuitt]

Hey Nerdy. Thanks for sharing your beautiful brain. I have a very similar environment to you ( 22.04 is different ) but sadly I can't run this and have installed it 3 times but keep getting segmentation faults. Any tips hugely appreciated as its late here in NZ. error directly after entering prompt == I oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
Segmentation fault (core dumped)