search

SUDO-AI-3D / zero123plus

Code repository for Zero123++: a Single Image to Consistent Multi-view Diffusion Base Model.
Apache License 2.0
1.56k stars 108 forks source link

Can I get train code? #76

Open Suprhimp opened 2 months ago

Suprhimp commented 2 months ago

Hi, first of all thank you about this cool model and method.

I want to make an pipeline similar with your model but I want to make single image input to single image output. For image generation that best camera view.

I think I can make dataset for that but I want to ask that how much do you need to train like your model. Also I want to train model that using depth and reference only controlnet. But I don't have any idea to start with.

Can I get some hint from the train code that you've made?

Thanks!

eliphatfs commented 2 weeks ago

Though not official, you can try the finetune code from https://github.com/TencentARC/InstantMesh