deep-daze VS VQGAN-CLIP

it's not nessecary to specify the model with the tool he is using also known as deep daze

(Added Mar. 15, 2021) deep-daze Fourier Feature Map - Colaboratory by afiaka87. Uses SIREN to generate images. Reference. Reddit post.

(Added Feb. 5, 2021) Deep Daze - Colaboratory by lucidrains. Uses SIREN to generate images. The GitHub repo has a local machine version. GitHub. Notebook copy by levindabhi.

Some people are wondering what I used to make it. I used the create_story feature of https://github.com/lucidrains/deep-daze. I then set the framerate such that it lined up with the length of the song and that was it :)

In case you have some damn good gamer-level GPU, you can try this FOSS alternative locally instead: https://github.com/lucidrains/deep-daze

If you're technologically savvy someone created a similar thing in Python, that runs locally on your computer. https://github.com/lucidrains/deep-daze

2. Deep Daze

If you have a pc head over to https://github.com/lucidrains/deep-daze and they tell you how to

I've screwed with many text-to-image models over the past couple of years, and I found that while I currently enjoy Stable Diffusion's coherency, I have a soft spot for the ImageNet model used by default for VQGAN+CLIP. It easily approaches the uncanny valley when generating people or animals, but makes for great abstract backgrounds and wallpapers. I already have nostalgia for generating images with it on my CPU overnight.

For someone only tangentially familiar with this space, how is this different than e.g. https://github.com/nerdyrodent/VQGAN-CLIP which you can also run at home? Is it the quality of the generated images?

I don't have any of the DALL-Es but I do have a couple from github [1], [2] which gave these outputs[3]

[1] https://github.com/nerdyrodent/VQGAN-CLIP

If you want to try your hand at buggering your whole system - try playing with AI image generation as it uses all possible computer assets :D . There is a lot of forms and installations for those but I VQGANs from github the easiest. Problem is that some require familarity with shell, python and in some cases - you need to enable the Linux subsystem in Windows (is it called a subsystem? it is not exactly a VM). This one is the easiest to install out of all I tried. But I liked the results of Pixray most but I wrecked it. I use this one nowadays.

There are various implementations floating around the web which combine VQGAN and CLIP to generate AI artwork (Google Colab or locally using Python). See this list for a more comprehensive overview of projects using VQGAN+CLIP.

Not for mine specifically but it’s all based on nerdyrodent’s VQGAN+CLIP with Real-ESRGAN

Have you checked out https://github.com/nerdyrodent/VQGAN-CLIP? It has options for input and target images that seem like they might be helpful.

Glad to hear that! I think dribnet has done more than just change default settings. It's based upon nerdyrodent's VQGAN+CLIP code, which I believe has no associated Colab notebook currently.