Stable Cascade img2img in ComfyUI

  Рет қаралды 1,791

FiveBelowFiveUK

FiveBelowFiveUK

Күн бұрын

Update: full Checkpoints ComfyUI now added, check here:
• Stable Cascade Comfy C...
Now img2img was released we can start to do some really interesting things, watch this space for more workflows and explainers !
effnet.safetensors encoder: (place inside comfyui/models/vae/)
huggingface.co/stabilityai/st...
Using ComfyUI, we can now take a look at Stability's latest Image Generation model, Stable Cascade. A three stage model, it can produce high resolution images with good alignment very quickly compared with competing baselines.
Article: civitai.com/articles/4161
ComfyUI Workflow: civitai.com/models/310409
You can download all the models inside ComfyUI Manager, be sure to update all your custom nodes in addition to ComfyUI. Consider disabling unused custom nodes if you have trouble.
discord: / discord
www.fivebelowfive.uk
stable diffusion cascade
stable diffusion lora training
comfyui nodes explained
comfyui video generation
comfyui tutorial 2024
best comfyui workflows
comfyui image to image
comfyui checkpoints
civitai stable diffusion tutorial

Пікірлер: 19
@AllanMcKay
@AllanMcKay 5 ай бұрын
Thanks for the great video! Keep up the awesome work
@davidbell6101
@davidbell6101 5 ай бұрын
thank you.
@juanchogarzonmiranda
@juanchogarzonmiranda 5 ай бұрын
Thanks!
@edmundoangulo2043
@edmundoangulo2043 5 ай бұрын
Great video! First img2img on KZfaq for Cascade from what I can tell! Has anyone tried to use the ControlNet models Stability AI provides for Cascade? Where do we put them - after Stage C or B. I can't get them to work...
@FiveBelowFiveUK
@FiveBelowFiveUK 5 ай бұрын
I have tried using Canny model, but i think there are some extra pieces we need to actually use it. Someone told me that the examples are incomplete. When it lands, there will be a new workflow and video ;)
@RonnieMirands
@RonnieMirands 5 ай бұрын
Can you help me please? My comfyui dont find these missing nodes: EmptyLatent and Conditioning. Searching doesnt return any results. Is there way to manually install this? Can find it. Thanks a lot in advance!
@FiveBelowFiveUK
@FiveBelowFiveUK 5 ай бұрын
try updating the comfyUI, i released a collection of workflows with good descriptions on Civitai, with basic workflows that only used built in nodes for the latest version. Some are more complicated, but there were some new dedicated nodes for Cascade. Which workflow are you using? they all have a name and version number to help
@vladyslav9098
@vladyslav9098 5 ай бұрын
Is it possible to add m2m video through this workflow?
@FiveBelowFiveUK
@FiveBelowFiveUK 5 ай бұрын
the beauty of comfy is we can try to find out, my advice is to try adding it and if it doesn't work, try a different way. Anything that works with WurstchenV2 should work with Cascade but it's still week one and a lot has changed already!
@andrevianaa2
@andrevianaa2 5 ай бұрын
Nice video. Silly question here, the workflow v18 still works for txt2img? im a newbie in this AI thing
@FiveBelowFiveUK
@FiveBelowFiveUK 5 ай бұрын
Yes the older v10--v18 are still available, they use a more complicated UNET & CLIP model loading setup, still works great, this new 2 checkpoint method is aiming to simplify it for people wanting to get started. as far as i can tell, the outputs are the same when comparing.
@andrevianaa2
@andrevianaa2 5 ай бұрын
@@FiveBelowFiveUK nice, thank you
@aksi221
@aksi221 5 ай бұрын
I am trying to get into this and learn more. is it possible to insert an image of an object without a background and mask ithe object in some way with an alpha and generate the background and not make any changes with the object itself
@turboscreenshooter
@turboscreenshooter 5 ай бұрын
Masks.
@FiveBelowFiveUK
@FiveBelowFiveUK 5 ай бұрын
OK so, Alpha channel is supported in Comfy, so it might be that you need to play with masking or attention masking. Another way is to use segmentation to "remove everything" around the subject. For example: You can load image into segmentation, then take that image and convert it to a mask, then you can take the mask into the attention for IPadaptor which also takes the original image at the same time. This will prevent any changes outside the mask. I will be covering things like this in future, so please come back and check :D
@fossil98
@fossil98 5 ай бұрын
I cannot find this on search engine, what exactly is 'shift' and where is it documented?
@FiveBelowFiveUK
@FiveBelowFiveUK 5 ай бұрын
discussion can be found here: github.com/comfyanonymous/ComfyUI/issues/2785 code/commit: github.com/comfyanonymous/ComfyUI/commit/8b60d33bb7ce969a53fc5e25bfa0e2dca7a17b23 that's the best i can do :)
@fossil98
@fossil98 5 ай бұрын
@@FiveBelowFiveUK Thanks :)
Stable Cascade in ComfyUI - Fully Supported!
7:04
FiveBelowFiveUK
Рет қаралды 9 М.
Real time Stable Diffusion in TouchDesigner
8:12
VJ SCHOOL
Рет қаралды 8 М.
Mom's Unique Approach to Teaching Kids Hygiene #shorts
00:16
Fabiosa Stories
Рет қаралды 37 МЛН
ВОДА В СОЛО
00:20
⚡️КАН АНДРЕЙ⚡️
Рет қаралды 35 МЛН
Foda FLUX1 Pack V3 | Auto Prompt Enhancer & No Easy Use Versions #comfyui
25:15
Easy Stable Cascade image to image & upscale with ComfyUI
11:54
Code & bird
Рет қаралды 2,2 М.
Advanced AI Projection Mapping Workflow
25:04
reflekkt
Рет қаралды 656
Stable Cascade Comfy Checkpoint Update
11:45
FiveBelowFiveUK
Рет қаралды 3,7 М.
Flux Completely Destroys Stable Diffusion 3! The New Champion
11:03
All Your Tech AI
Рет қаралды 20 М.
BASIC - All you need to know about Stable Diffusion (Mac & Windows)
8:30
Stable Cascade workflow in ComfyUI
8:37
Neuron
Рет қаралды 3 М.
ComfyUI for Everything (other than stable diffusion)
32:53
Design Input
Рет қаралды 31 М.