V2 - Advanced weighting and various improvements
Highly recommend just using the new workflow (included .zip or in example workflows on the github)
Installation:
You need DiffSynth-Studio FIRST:
CLICK THIS TO GO TO MY GITHUB README FOR INSTALLATION INSTRUCTIONS FOR YOUR SETUP
(But I already installed):
Please go to the link above anyway and confirm THAT is how you installed.
(Optional) Before Starting:
You can use the built in modelscope download to grab the files for you automatically though you can get them much faster manually with huggingface if you know what you are doing (and where to put them) Otherwise, just be patient, it only needs them once.
https://github.com/by-ae/ae-in-workflow
Description
FAQ
Comments (5)
WHAT IS THIS?
It is a couple of workflows using my new custom nodes to go from a folder of images to a lora to an image in a single pass within ComfyUI.
ISSUES?
Read the info carefully. Do you have DiffSynth-Studio cloned and installed within the same environment as ComfyUI? Do you have my simple workflow loaded? Is your batch size set low if you have low VRAM?
Comment any errors in the ComfyUI log/terminal and I'll help or explain.
The image selector node simply gets my comfyui stuck to the point where I can't even ctrl-c it, only kill the process. It also doesn't seem to be doing anything productive, no GPU, CPU, network or RAM usage, just dead stuck.
Edit: Managed to load the images with a different node, now it's saying diffsynth-studio is not installed, double checked and I certainly have diffsynth 2.0.3 listed in pip list for my comfy envionment.
@fxvoid How are you installing it into your comfyui environment?
Both the custom node and the DiffSynth-Studio.
That is the correct version but the code is very simple for that check.
Just attempts to import and raises that error if it fails, so I suspect env.
Check out the github, it has a more detailed installation run down for different ComfyUI installations now (specifically for DiffSynth-Studio as it is the painful one).
@fxvoid I am also curious about the freezing on the image selector. It's pygame based. So pretty battle tested lib, nothing fancy. That's why I'm checking how you are installing the custom node also.
Let me know and I'll see if this leads to some improvement in the readme.
@by_ae Thanks for the replies, I have to admit I tried to install it via comfyui-manager but it said I can't install it this way due to permissions level or whatnot, so I decided to just git clone the repo to custom_nodes and run pip install -r requirements.txt, worked just fine with other node packs before. I installed the DiffSynth studio by activating my pyenv comfy environment, cloning the diffsynth repo and running pip install -e . from there. It could quite possibly be just my environment, in which case I apologize for wasting your time. Regarding the image selector, I can suggest making it javascript-based instead so it can run directly inside the browser (very similar functionality already implemented here: https://github.com/heyburns/image-chooser-classic)