Flux Dev (Black-Forest-Labs) - Installation Tutorial

Black Forest Labs has gotten the world excited with it's recent release of their Flux models. And the quality is so good that many are saying it marks the official end of Stable Diffusion. And now you can try it yourself!

In this tutorial we walk you through literally EVERY step of installing this model on your computer. The installation includes the ComfyUI app, detailed directions on downloading the models, and a basic tutorial on how to start your generation journey! Noob friendly

If you would prefer to get going fast, we have a one-click-installer on our Patreon site that does all the work for you. Get up and going in mere minutes!

------ Helpful Video Tutorials -----

[pilotdawn]

Excellent video demo - extremely helpful

[CiY]

this channel is amazing, I love your hat, I love your style. even though I have experience in programming, the way you explain things doesn't make me feel stupid, it makes it all easier.

[TalhaKarakoyunlu]

Thanks for the informative video, I was trying to tackle this by myself for the past week and let me just say that it is not easy when you are new to the topic. Also, do you have a discord server? If so I would like to join it.

[VintageForYou]

This is an amazing AI tool I installed this the other day using on ComfyUI, and now I am working to do some good CLIP Text Prompts to produce some great images.😁👍

[trypovios]

You forgot to mention that Conda should be pre-installed on the system before you can install Nvidia cudatoolkit into the virtual environment. Without that prerequisite, these instructions are not going to work for most people.

You do not have to have Conda to install the toolkit; it can also be done via "pip install nvidia-cuda-runtime-cu11 nvidia-cuda-nvrtc-cu11" which will install toolkit 11.8.

[simuman]

Anybody following along with this tutorial and getting an issue when generating a image about "Numpy is not available". First check what version of Numpy you have by stopping the server and trying to install with (no quotes) "pip install numpy". If you get a message like "Already installed version 2.0.1, then run this command "pip install numpy==1.26.4". This should fix it.

[orsiswed]

What if your system is Windows10 and the GUI is an AMD ?