If you have an old GPU with 4Gb of Vram, I highly recommend to use ComfyUI as it's much faster than Automatic1111 and InvokeAI. Also ConfyUI is very powerful really easy to install. Don't waste your time and energy try ComfyUI: github.com/comfyanonymous/ComfyUI
I am running ComfyUI inside automatic11111, and also running automatic itself and all its extensions on 4GB. ComfyUI is very easy to install yes but it is not easy to understand. Automatic is easier to understand but tricky to install (I'll make a list of common errors and solution in a week or 2), moreover it allows you to access to incredibly useful stuff like Deforum, ControlNet, text to video, depth maps, pose estimation ,vector studio (to make vectors directly from prompts), prompt generator, artist database, training your own models, 360 diffusion amongst others. Also has plugins to directly incorporate photoshop and ebsynth. The main goal of recommending automatic is to have complete control on most AI vision related stuff for creatives. If one merely wants to use Stable Diffusion for images and play around with modular workflow then you can use ComfyUI (it doesnot even support videos yet). Perhaps AInodes will work as an inbetween but it has lots of bugs at the moment.
It looks like there is an EASY way to install many kinds of AI stuff, Pinokio is a browser that lets you install, run, and programmatically control ANY application, automatically. pinokio.computer/
If you have an old GPU with 4Gb of Vram, I highly recommend to use ComfyUI as it's much faster than Automatic1111 and InvokeAI. Also ConfyUI is very powerful really easy to install. Don't waste your time and energy try ComfyUI: github.com/comfyanonymous/ComfyUI
I am running ComfyUI inside automatic11111, and also running automatic itself and all its extensions on 4GB. ComfyUI is very easy to install yes but it is not easy to understand. Automatic is easier to understand but tricky to install (I'll make a list of common errors and solution in a week or 2), moreover it allows you to access to incredibly useful stuff like Deforum, ControlNet, text to video, depth maps, pose estimation ,vector studio (to make vectors directly from prompts), prompt generator, artist database, training your own models, 360 diffusion amongst others. Also has plugins to directly incorporate photoshop and ebsynth. The main goal of recommending automatic is to have complete control on most AI vision related stuff for creatives. If one merely wants to use Stable Diffusion for images and play around with modular workflow then you can use ComfyUI (it doesnot even support videos yet). Perhaps AInodes will work as an inbetween but it has lots of bugs at the moment.
It looks like there is an EASY way to install many kinds of AI stuff, Pinokio is a browser that lets you install, run, and programmatically control ANY application, automatically. pinokio.computer/
youtu.be/Gunh6VoEJuU
It's possible to use a sequence of images as img2img with comfyUI, there are tons of extra nodes you can find