Adjusted model and diffusion configs do not work

[LucaFilitz]

Basically, I want to increase the number of points which are generated by text2pointcloud. I know that I can't just increase num_points in the sampler. So, I adjusted the base and upsample model and diffusion configs to a higher point number. My problem is, that CLIP seems to have a problem with that. (I get an Error when loading state_dict for CLIPImagePointDiffusionTransformer because it does not match the new size of my adjusted models)

Is this a known problem/ is it just not possible to increate point quantity?