Stable Diffusion 3 via API in comfyUI - Stable Diffusion Experimental
Ғылым және технология
Stable Diffusion 3 has been released to use via API key and today we'll take a look at how we can run it inside of comfyUI with Stablity AI's official nodes.
Disclaimer: since SD3 has been released for use via API key only right now, it's not free. New accounts have 25 free credits, which should be enough for 2-3 generations.
If you like my contributions, you can buy me a coffee here: ko-fi.com/risunobushi
Links:
- SAI SD3 nodes github: github.com/Stability-AI/Comfy...
- Stability AI SD3 webpage: platform.stability.ai/docs/ap...
- Stability AI API keys: platform.stability.ai/account...
Timestamps:
00:00 - Intro
01:04 - Installing Stability AI SD3 nodes
04:49 - Getting the API Keys and free credits
05:52 - Core and SD3
10:44 - Upscaler
11:56 - Outpainting
14:38 - Inpainting
16:28 - Search and Replace
18:49 - Outro
#sd3 #stablediffusion3 #stablediffusion #stablediffusiontutorial #ai #generativeai #generativeart #comfyui #comfyuitutorial #risunobushi_ai #sdxl #sd #risunobushi #andreabaioni
Пікірлер: 11
Leave some prompts here and I'll run them using the remaining credits I have! (until I run out)
"Could it better? Yes. Is it funny? Also Yes". 😆
Good walk through. Thanks
great videos man
your key revealed
@risunobushi_ai
Ай бұрын
Yep, but I changed it and got a new one before uploading the video, the “I’m not falling for that” was a joke :)
@PixelArt_YW
Ай бұрын
@@risunobushi_ai Thanks for your video, It's helpful.
They've defeated the entire point of Stable Diffusion as a company
@risunobushi_ai
Ай бұрын
Base models have always been released as usable with API keys only first, and a few weeks later they have always released the checkpoints. It was the same for SDXL, it hasn’t changed
@TheBestgoku
Ай бұрын
They will release full weights soon, its too big to run on most local PC's anyways.
@generichuman_
Ай бұрын
Man, people get really upset when they stop getting free shit for nothing... Also relax, the weights are coming soon, the same thing happened with SDXL, first an api, then the model is put out.