[Curious] Image to Mesh

[matbee-eth]

We could seemingly train the dataset on an image:vertexes pairing to get what would be essentially the equivalent of a depth-map to mesh language model?

I'd be very interested in the training code when available! Would be a fun experiment.

[thuwzy]

Hi! We have actually thought about training a image to mesh version. The idea is neat and likely to work. However, owing to the computational resource we haven't tried that yet.

As for the training code, we use LLaMA-Factory, which is a very easy-to-use codebase for LLM traninig.