Update (3/11/23): Testing Nolan Aatama's newest sd-1click-colab for anything4.5 launched in just about 5 minutes (nice!) However they have added/changed most of the code. If if ever have the strength I may adjust to it, if there's anything crazy improved then I will add/change to it if I find out about it.
Update (4/18/23): Testing Nolan's latest script took around 9 minutes to install and launch a public URL for the webui (I acceidenlty stopped the time midway so I don't have an exact time) Might try update to Nolan's latest changes, it looks a lot more feasable now.
Update (4/18/23): Finally got around to getting a Kaggle script. Kaggle gives you a bit more freedom than Google for free, especially with it's 30 free GPU hours per week (that's around 4.28 hours of usage per day!). AND YOU GET PERSISTANT STORAGE (15 GB)!!!!! It's really just camenduru's script with a fix for the image browser to work with old gradio versions. Camenduru hasn't updated the script in a bit and I haven't figured out how to update Gradio without breaking other stuff, so, yeah. Also included a cell to easily launch the web-ui after installing. Will add more later if i'm up for it.
Update (7/28/23): I am now in possession of a slightly more powerful laptop, if I keep this laptop for long enough and see that installing a1111 locally gives enough speed and preformance for generations in a reasonable size, I will archive this repository. If this doesn't happen to be the case, I will hopefully revisit this script eventually, but I hold no promises.
(Using Google colab) First cell startup time - around 4:50 sec. (Web-UI fully loaded) (using current Google script that is outdated from upstream)
(Using Google colab) Second cell startup time - aprox. 1:33.52 sec. (Web-UI restarted, Google Drive mounted and option files from Google Drive loaded) (using current Google script that is outdated from upstream)
Estimated total startup time for most users on free Google Colab account: ~ 5.8 - 7 minutes (using current Google script that is outdated from upstream)
These are just edits/modifications of people's script, I just made some modifications to make it easier to understand various things such as changing the model and vae, adding extensions and also adding in support for loading options (option loading limited to Google Drive currently).
The script is already setup with Anything-V4.5 model and Anything-V4.0 vae, but can be changed to any by replacing the links and file names.
IMPORTANT Only Google Colab and Kaggle scripts currently. Will update with Paperspace script if i'm ever able to claim a free GPU or decide to subscribe for some reason (highly unlikely).
Google Colab Script: Changes from Nolan Aatama's original sd-1click-colab (currently edited from a outdated version, still need to update to latest version/upstream and add more extensions)
Added autocomplete booru extension
Added booru2prompt extension
Added booru tag autocompletion extension
Added Civitai browser extension, forked version by Etherealxx
Added easy way to mount drive and replace option files to mitigate sessions destroying all resources after being ended
Added many comments including explantions on changing/adding in more models and VAEs, adding extensions, as well as a super short tutorial on setting up the options replacement with your Google Drive (Only neccesary once!)
Sampling steps won't change to Google Drive saved options when replacing with ui-options.json from your own PC
Comments for explaining how to add hypernetworks and embeddings with web links (the same way you can add models and vae's)
Create Kaggle version (current Google colab version spits out some errors even after replacing file storage directorys to Kaggle's and removing any Google drive functions to take from another service) Ensure Kaggle script stability and full functionality (whatever functionality available in the A1111 WebUI that is to the extent of my knowledge, as well as whatever extensions I know and use)
Create Paperspace version
Retest current Google version to ensure stability after updating everything
Create mobile version?
As iterated from the About this script section of this README.md, I do not own any of the original code made for these scripts/colabs nor did I contribute to the original scripts.
I am not a python developer, I am not a C++ developer, I just try to configure things to the extent of my knowledge and best ability for myself and anyone else who would like a better pre-configured expirence with little to no hassle. Any changes I make to the original scripts/colabs by thier original creators are only changes that ANYONE can do, with absolutely 0 coding knowledge.
All respective credits go to (will update as needed):
Nolan Aatama for their immense sd-1click-colab scripts/colabs and amazing resource of tutorials on their (Youtube)[https://www.youtube.com/@nolanaatama].
camenduru for their work on (will add later) and all developers of the extensions I added, the people who created the models I added, the people behind Civitai, and anyone else I may have forgot to mention.
They are all the real people to spread your thanks for these scripts/colabs, as it would not be possible without their work.