Can't run the app

[Dulmen]

it gives me this error when i launch webui

RuntimeError: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 2.00 GiB total capacity; 1.73 GiB already allocated; 0 bytes free; 1.77 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

What i have to do? i'm a novice.

[biohack78]

2GB of VRAM isn't enough to run SD. I'd recommend using a colab like this one: https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb

[Dulmen]

can't understand how it works. Can you explain me? i'm i novice

[biohack78]

Sure thing

Step 1: Make a huggingface.co account and navigate to: https://huggingface.co/settings/tokens

• Click on 'New Token' and give it a name (the drop down should be set to read). We will need to use that token soon.
• Navigate to: https://huggingface.co/CompVis/stable-diffusion-v1-4 and accept the terms and conditions.

Step 2: Head to: https://colab.research.google.com/github/TheLastBen/fast-stable-diffusion/blob/main/fast_stable_diffusion_AUTOMATIC1111.ipynb

• On the top right, click on the button that says 'Connect'. That will allocate you a server for free.
  
• These buttons run the code in each box. Start with the first one. It will probably ask for permissions to use your google drive, that's the storage for the repo, model and images you generate (I think all-up it's around 6gb or something like that). Once you see a green tick next to that cell, you can move onto the next one.
  
• Paste your token from earlier into this box and hit enter
  
• Some of the cells takes a few min to install/setup. Just wait till it finishes.
• At the last cell, it should give you a gradio link, follow that and you should get a webui to the server.

Happy generating