sample workflow please

#4
by jartoph - opened

i am having a hard time figuring out how to get it working, would be nice thanks

@jartoph I made one for myself and can share. It does use my custom nodes because I had to change a few things in order for the nodes to work in my video builder. So just install my custom nodes. upload a ref image of someone and enter a prompt then run it. It does use the turbo lora though but it just makes it MUCH faster. 8 seconds at 1080P only takes 455 seconds on a 5090.
workflow

the workflow just be embedded in the image. If it does not work let me know.

Just noticed ping pong is on so just turn that off in the video combine node!

@vrgamedevgirl84

It does use my custom nodes because I had to change a few things in order for the nodes to work in my video builder. So just install my custom nodes.

You're a gem for sharing! Your workflow uses two nodes that aren't part of your Comfy Nodes pack (and aren't found in the manager).
TE-Speed-MiniMaxH3-OSS - which seems like a complex accelerator. Not sure if compatible with all hardware?
comfyui-minimax-h3-audio-T8 - which seems even more complex and I have no idea what it does. Something to do with voice audio, face detection, and character separation?

Assuming neither are needed, but curious what your usage & thoughts are about those two extras.
Also, I just found your Comfy-Krea2-MLX nodes via the manager, but the github repo is gone. Is it no longer available?

Thanks again, and for all your contributions!

@vrgamedevgirl84

It does use my custom nodes because I had to change a few things in order for the nodes to work in my video builder. So just install my custom nodes.

You're a gem for sharing! Your workflow uses two nodes that aren't part of your Comfy Nodes pack (and aren't found in the manager).
TE-Speed-MiniMaxH3-OSS - which seems like a complex accelerator. Not sure if compatible with all hardware?
comfyui-minimax-h3-audio-T8 - which seems even more complex and I have no idea what it does. Something to do with voice audio, face detection, and character separation?

Assuming neither are needed, but curious what your usage & thoughts are about those two extras.
Also, I just found your Comfy-Krea2-MLX nodes via the manager, but the github repo is gone. Is it no longer available?

Thanks again, and for all your contributions!

these are not my custom nodes. They may be part of some of my workflow's but I did not make them. You should be able to do "find missing custom nodes" to find them.

fyi in the example workflow that now exists there is a error. The height from the size selector is connected to length not to height.
gave me a oom.
grafik

fyi in the example workflow that now exists there is a error. The height from the size selector is connected to length not to height.
gave me a oom.
grafik

did my workflow seem to work after installing my custom nodes?

i posted it as a full post and it has been fixed within a few hours. all is well now. I did have a allocation on device error ( basically out of memory) error at the very end, on the vae decode on my 16gb vram pc. length+resolution was too much for it. Using a vae decode tiled node instead (just connect it the same way) solved that problem.
Other than that it works well.

image

Sign up or log in to comment