I'm a complete newbie to Linux—just got started less than a week ago. I've been following a detailed YouTube tutorial about using ComfyUI, which has a ton of helpful videos. However, I'm currently stuck on episode 11. The host mentions that I need Windows and shows a command to run in CMD: "python_embeded\python.exe -m pip install -r ComfyUI\custom_nodes\ComfyUI-Florence2\requirements.txt". I've checked the file path, and it seems correct. So far, everything from the tutorial has transferred well to Linux, but I'm nervous about opening a terminal and running that command. Are my concerns unnecessary, or could I accidentally mess something up?
1 Answer
To get ComfyUI running on Linux, just follow these steps:
1. Install Git: You can do this by running `sudo apt install git` if you're on an Ubuntu-based distro.
2. Make sure you have Python installed, and it's a good idea to use PyEnv to manage different versions of Python.
3. If you have an Nvidia GPU, make sure to install both Cuda and cuDNN (I recommend Cuda 12.4).
4. Clone the Comfy repo using Git.
5. Open a terminal in the folder where it downloaded.
6. Create a virtual environment with the Python version you want to use.
7. Activate that virtual environment.
8. Check the Comfy Repo's Readme for a pre-install script for your graphics card, run that in the terminal.
9. After that, type `pip install -r requirements.txt` in the activated terminal.
10. To install the ComfyUI frontend package, run `pip install comfyui-frontend-package --upgrade`.
11. Finally, launch the app by running `python main.py` in the terminal.
That's not exactly what I was asking. I already have ComfyUI installed, but I need to install Florence2 through it, and it requires terminal commands that are meant for Windows. Is there a Linux version that won't mess things up?