Fooocus UI - Run in Colab directly from Google Drive
ฝัง
- เผยแพร่เมื่อ 18 ธ.ค. 2023
- Fooocus UI - Stable Diffusion
Link to Google Colab: colab.research.google.com/dri...
It's annoying that when using Stable diffusion in Google Colab you need to re-install the Fooocus library and any models every time you load it.
This method allows you to install the Fooocus repo and any models once in Google drive, then load them into your Colab on each run.
It also means that files generated during the session are stored in Google Drive by default.
You my friend are an absolute legend for making this, super helpful, thanks so much!
Wow. A tutorial that just shows how it is done. Direct and to the point! Thank you! (Subbed)
I couldn't thank you enough! Brilliant!!!
Thank you - I was looking for this.
Omg, thank you so much! You are really legend!
Love you bro! Hope you get 100M subs😍
thank you very much! But would it be possible to install the "ReActor" extension? Why is the face swapping feature really bad in Fooocus!
this error Error checking version for torchsde: No package metadata was found for torchsde
Installing requirements
Anyone having trouble downloading models? When I try to download it says it already exists but when I check my checkpoints folder, they aren’t there
Hi, can you tell me how to make a dark theme
make more videods man ur fureaking amazing!!!!
Thank you for sharing. Working well. Except that the html on the outputs is converted and stored in a google doc, losing the interest of having the initial html. Do you have a solution to get it right?
I'm always thankful. I run fooocus with Colab. Is there a way to install an extension? Also, when doing faceswap, how can I make it look as similar to the person as possible?
Hello, thank you for making this. How can I use a different model? I tried entering the link to different models, but it wasnt working.
Thanks alot bro!
Thanks for this method! Is there a way to download checkpoints and LORAs directly from civitai to Google Drive? You made a way to download a custom model, but what if I want to download several at once? And not only checkpoints but also LORAs.
have you found out how to do it?
Thank you so much for this! Would we need a collab pro subscription for this?
How do I use other features such as inpaint, outpaint, faceswap etc???
Can we do a face swap here as if we do it directly google colab stops us from doing it...?
Can you add the link of the model that you used into the description? or should i get from CivitAI?
it's NOT lifetime and NOT free. after a few runs, google colab free can't reconnect anymore, is ther an alternative
how to install controlnet in fooocus and use in Colab?
When its run and put the propmt its stop and tell that there is disconnect error plllllz help
thanks for the great tutorial boss, I am new to this, could you provide some guidance on where to get the most up to date juggernaut xl link to install it thru your code? TY 1000
Where do i get the Model Link for juggernaut ?
Thanks for the tutorial. But don't we risk reducing the storage space of Google Drive, especially if we add models?
please make a tutorial for hugging face diffusers
Is it possible to do exactly the same but instead of using Google Drive, use Microsoft Onedrive? There I have 1 TB and I could upload all the models I wanted.
very helpful thank you! Google collab is kicking me out for inactivity when I use stable diffusion, do you have any idea why?
Are you using pro or free? Colab is a shared code environment so they try to clear up anything not being used. I imagine free tier even more!
@k5c5 thank you now that I upgraded it doesn't kick me out anymore
@@user-lc3ip4cw8h yeah for the ~$9/10 per month I just pay it 😄
what about NVIDIA DRIVER ?... do i have to download?
For some reason I am not able to run it locally, only trough gradio. Any idea why?
Very simple: Locally mean "installation was done on your computer" but you install on colab.
Hi so how are we supposed to get the link to install the model
You can use any model you like, that's the idea. The default is juggernautXL-v6: huggingface.co/lllyasviel/fav_models/resolve/main/fav/juggernautXL_version6Rundiffusion.safetensors
you didn't post the model link here why
Why swap face error generate
how to load more that 1 checkpoint, loras and vaes ?
For new models/ Check points I copy the download url and paste it in at the models link, and then run the Model download/Load code again. Then check in fooocus model folder at checkpoints for new model/checkpoint. Also want to now how to add Lora etc. Other video said you can download Loras to pc and drag and drop downloaded file in the Fooocus Loras folder and it will upload. Did not try it yet.
How to bypass the GPU limitations 😢 , I can’t create more than 4 images, it’s keep getting exhausted
Did you find any solution for that ?
i want to add shared drive.
what should i use in the blanks.
drive name or id?
same question. need to answer ASAP
Hi, is it possible to use 1.5 model instead of SDXL ?
In the step that you specify the model, you can use the hugging face or civitai link to any model you like 👍
@@k5c5 okay thanks for reply +1 sub :)
How to add your own photo with Virtual Influencer
NotImplementedError: You have selected base model other than SDXL. This is not supported yet.
Annoyingly Fooocus only supports models based from SDXL, I'm trying out Automatic1111 and ComfyUI as alternatives.
This is the fooocus issue in github: github.com/lllyasviel/Fooocus/issues/1568
it gives me error when producing an image. it just says Error and produces nothing.
bcs of collabs gpu stock , u should buy premium collab or restart foocus again.
@@sothe283 yeah it's sad. But i don't wanna pay effin google, or to de-anonymise myself. Which are what attracted me to this method in the first place. But thanks
What is the solution of disconnections
are you using a pro account? I found paying the monthly fee was worthwhile
@@k5c5 nah, I'm broke 😅