I have installed it, how do I run it?

[Lodjuret]

Run the python file called main.py

https://github.com/comfyanonymous/ComfyUI#running

in terminal do 'python main.py' when you are in the ComfyUI directory

[king-open]

How to choose a model?

[jokandre]

@king-open at first you will have no model, at least was my experience. As soon as you click on that "v1-5-pruned" it will become null. After making sure you have a model under the folder models/checkpoints, click the refresh button on the right or restart the server, then refresh your browser. The names of the models on the folder will appear as soon as you click on ckpt_name

[king-open]

In my "models/checkpoints" folder, there is a model, and I have also restarted the server. However, the names of the models in the folder are not appearing. Oh!

[jokandre]

I would recommend downloading a model and pasting it there yourself, that's what I did when I installed ComfyUI last week.

[king-open]

@jokandre I want to know where to download the models. Please provide the link.

[jokandre]

You can download checkpoints from civitai.com or hugginface. A Stable diffusion checkpoint will be tagged SD, usually 1.5 is fine, around 1.6Gb. There are also SDLX checkpoints the files are about 5.6Gb.

I am using SD 1.6 checkpoints and loras to learn ComfyUI for now, that easier for my RTX3060 laptop, 4G VRAM is more than enough for a simple workflow

[king-open]

We have encountered this issue

[king-open]

After debugging, I found there were some minor issues. Can you help me with this