So Runpod isn't an option either then?
So Runpod isn't an option either then?
![Digital [Starburst]](https://cdn.discordapp.com/avatars/220961909447458817/8ecbbb3132fc6e44312f3d6d2ae3df69.webp?size=40)
DRunpod is what I paid for actually if I remember correctly
![Digital [Starburst]](https://cdn.discordapp.com/avatars/220961909447458817/8ecbbb3132fc6e44312f3d6d2ae3df69.webp?size=16)
Frunpod is the correct option
Fvery cheap
DOkay. And I should use the Community option then.
DThanks.
Fyep
Fyou can use community
Fworks great
Ithanks so much 
DSo the project "mharleman-40925" is actually the name of the project I created in Kaggle. Google Cloud gives you $300.00 for 90 days to try out the service, and sets your first project as billable if it uses any services. To fix it, you have to go into the management console and disable billing for any projects you've opened. I never clicked on the button in Kaggle that told it to upgrade the project to Google. It all happened without doing anything.
DBuyer beware.
DI think since Kaggle is connected to Google it checks to see if you have a paid Cloud account. If you do, it automatically does this without telling you apparently assuming you know.
DIf I had enabled billing on the new project, I wonder if Kaggle would've worked?
DLet's try an experiment shall we?
Flet me know
DGoogle Cloud isn't a requirement. ngrok's instructions confused me. It told me to open the tunnel from my machine not from my notebook. You are only allowed one endpoint at a time. Really, you only need one. I have it working fine now. Sorry for the confusion. I think you can use Google Cloud if you don't want to run the install every time you login to Kaggle, but I'm not sure.
DI have Kaggle working. You're telling me RunPod is better though, and pretty cheap. If that's the case before I get too much into Kaggle, I will learn RunPod.
DOf course, Kaggle doesn't get any cheaper.
DIf RunPod gives me advantages though, I would rather spend a little money for those advantages. Kaggle isn't exactly fast.
Frunpod is way better than any paid service
Fbut if you want free kaggle is best
DSo I have ComfyUI running on Colab. It's really slow, but it says it is connected to a T-4 GPU. I get errors saying that the version of Pytorch is 12.0 but the version of Cuda is 11.8 and they aren't compatible which leads me to believe that's why it is slow. The other possibility is that I need a pod running. I don't know for certain how it works yet. I added a piece of code to install the ComfyUI_Manager to it, but even though the custom_node is present it isn't showing up in the interface. Not sure what I did wrong there. I know enough linux to check directories and by using commands you already have in your code it installed.
Twhy this error keep on showing 'Nonetype' object has no attribute? while start render the stable diffusion
can anyone explain why and how so that i can rectify. pls help
can anyone explain why and how so that i can rectify. pls help
Fwith colab problem is we cant make venv
Ftry latest dev version of automatic1111
Tok sir
Tcan you share me a link pls
Ti cant find it in Git
TFopen your automatic1111 installation folder in cmd
Ftype
Fgit checkout dev
Ftype
Fgit pull
TOK
Tstill same i dont know what to do
Tsorry to bother you sir
Fah you are trying deforum
Fi havent used it yet
Fi would try to do a complete fresh install
Ti did twice
Fmaybe your used model is not supported
Ti took the model for civit.ai
Trev animated
Twhich model supports this dev automatic 1111
Teven youtube videos tutorial are not working
Fok i just noticed your error
Fyou have 2 GB having GPU
Fyou wont run it
Tso what can i do ?
