how do I create more than 130 iterations?

[AdamMady]

[KeatsHao]

It doesn't create 130 iterations actually. There are two parts: generating 1024 points and upsampling 3072 additional points. Each part has 64 default steps in karras_sample_progressive. Plus, in the end of them, they each yield the final prediction which makes the total number of yield 65. So the final "iterations number" is 130. You can check the code in point_e/diffusion/k_diffusion.py. And for your question, just change the default karras_steps in PointCloudSampler initialization.