diffusion-models-class icon indicating copy to clipboard operation
diffusion-models-class copied to clipboard

Unit1/01_introduction_to_diffusers.ipynb

Open yao-yu1999 opened this issue 2 years ago • 4 comments

This link doesn't seem right. Have you changed the address? https://github.com/huggingface/diffusers/raw/main/examples/unconditional_image_generation/train_unconditional.py Is this link the same as the one above? "https://github.com/huggingface/diffusers/blob/main/examples/unconditional_image_generation/train_unconditional.py" If yes, do you know what the error is and how to solve it?

yao-yu1999 avatar Feb 23 '23 06:02 yao-yu1999

It looks like the script is trying to import a library that isn't installed on your machine. Try pip install tensorboard - hopefully that will solve it

johnowhitaker avatar Feb 24 '23 06:02 johnowhitaker

Wow, problem solved, thanks!

yao-yu1999 avatar Feb 27 '23 01:02 yao-yu1999

In Accelerate script there is warning coming.

UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to the gradient layout contract, or that the param's strides changed since DDP was constructed. This is not an error, but may impair performance. grad.sizes() = [512, 1024, 1, 1], strides() = [1024, 1, 1024, 1024] bucket_view.sizes() = [512, 1024, 1, 1], strides() = [1024, 1, 1, 1] (Triggered internally at ../torch/csrc/distributed/c10d/reducer.cpp:323.).

Can this cause issue?I guess this is because of not using contiguous after permute somewhere in diffusers.

ayu1729 avatar May 18 '23 09:05 ayu1729

@ayu1729 Hi, did you find any solution your above mentioned warning?

igeor avatar Dec 19 '23 11:12 igeor