improved-diffusion

improved-diffusion copied to clipboard

I have implemented a diffusion probabilistic model for the first time using a cosine noise schedule, and I am finding that the variance/standard deviation of samples drawn from the reverse...

jamesheald

We are training a 3D diffusion model on medical data using this library and facing some instabilities during training. More in detail: The image quality stagnates/decreases after 80+ hours of...

simonasantamaria

After looking at the loss design, my understanding is to set --learn_sigma False during training. Is it correct?

neithen-Lu

Fail to resume model through mpi load_state_dict

2

hi, thanks for your awesome work, I have used it to do some interesting topic. But one thing still bother me, I cannot resume my model checkpoint when I do...

fangchuan

saving checkpoints frequency

1

Hello, Is there a possibility to observe samples let's say every 1K steps to monitor training ? Thank you

NicolasNerr

how to use ddim_reverse_sample?

2

I am trying to use the ddim_reverse_sample to get the noise used to generate the image. I simply replaced the ddim_sample with the ddim_reverse_sample in the ddim_sample_loop_progressive function, but it...

wrx33

How can I terminate the training?

2

This training seems to lack the concept of epochs and is trained batch by batch. How can I obtain the model with the best parameters? Could someone provide an answer?...

ykcheng9966

Query: Samples Files

1

I wanted to ask how to analyse (visualise and compare) the npz files from sampling.

KomputerMaster64

Training time

4

Does the training never stop if `r_anneal_steps` is not manually changed?

JesseWiers

p_mean_variance mean calculation

2

I was looking through the code to see how the paper was implemented, but I ran into an issue when looking at the part of the paper measuring the KL...

gmongaras