xaranex1

I have now new problem: [WARN] No checkpoint found, model randomly initialized. i did parameter --dmtet --init_with trial_image22/checkpoints/df.pth but it not start from that checkpoint

How can i used it in code? Can you write some example for jupyter notebook?

You think nn in models? So if i put nn instead of nerf it will be work?

I dont know how can i do it, jupyter notebooks examples are without torch.nn

I tried to modified code like this: xm = torch.nn.DataParallel(load_model('transmitter', device=device)) model = torch.nn.DataParallel(load_model('text300M', device=device)) but i have error: AttributeError: 'DataParallel' object has no attribute 'd_latent'

Is there some solution?

How can i use multiple GPUs ?

Hi, Kaedim3D has the option of an optimal 3D model result that can be printed on an FDM printer. For now, this is the only solution available.

yes same mistake for me but in colab it works fine..

I tried to change guidance_scale to 20 and karras_steps to 15000 but quality is same as before