denoising-diffusion-pytorch VS stable-diffusion

I’m trying to train a diffusion model using this code (https://github.com/lucidrains/denoising-diffusion-pytorch). My idea is to take a short audio segment, transform it into a spectrogram and train the model on these images then have it generate spectrograms then go back to audio. However the model requires square images. I cannot for the life of me figure out how to make a square spectrogram. Also is a regular spectrogram or a mel spectrogram better for this application?

Generally it's without weights, but MusicLM is also a WIP more mature implementations have descriptions on how to train them and follow ups on small scale/crowd-sourced experiments & research[1].

[1]: https://github.com/lucidrains/denoising-diffusion-pytorch

[1] https://colab.research.google.com/drive/1sjy9odlSSy0RBVgMTgP7s99NXsqglsUL?usp=sharing#scrollTo=KOYPSxPf_LL7 [2] https://github.com/lucidrains/denoising-diffusion-pytorch/blob/main/denoising_diffusion_pytorch/denoising_diffusion_pytorch.py

Sure. I am using: https://github.com/lucidrains/denoising-diffusion-pytorch

Lucidrains GitHub is always my go to repo for understandable paper implementations https://github.com/lucidrains/denoising-diffusion-pytorch

Thanks for the reply. Is there any suggestion if I wanted to train a model to generate half cat and half butterfly images what I should do? I git cloned the code from https://github.com/lucidrains/denoising-diffusion-pytorch and trained from scratch.

DDIM/DDPM are the same model to train, they only differ at inference time. To start I would recommend building from lucidrains' MIT licenced version (https://github.com/lucidrains/denoising-diffusion-pytorch). Just play around with the models until you gain an intuition.

Our codebase for the diffusion models builds heavily on OpenAI's ADM codebase , lucidrains, Stable Diffusion, Lightning and Hugging Face. Thanks for open-sourcing!

Once you understand these papers you can begin to understand Palette, and from there I would start with an open-source diffusion implementation like this one and then modify it to suit your needs!

Which is a 1-click install of Stable Diffusion with an alternative web interface. You can choose a different approach but this one is pretty simple and I am new to this stuff.

an invisible watermarking of the outputs, to help viewers identify the images as machine-generated.

It sounds like you just want to set up Stable Diffusion to run locally. I don't think your computer's specs will be able to do it. You need a graphics card with a decent amount of VRAM. Stable diffusion is in Python as is almost every AI open source project I've seen. If you can get your hands on a system with an Nvidia RTX card with as much VRAM as possible, you're in business. I have an RTX 3060 with 12 gigs of VRAM and I can run stable diffusion and a whole variety of open source LLMs as well as other projects like face swap, Roop, tortoise TTS, sadtalker, etc...

Use specific GPU on multi GPU systems · Issue #87 · CompVis/stable-diffusion · GitHub

I dealt with this in one of my posts. At least SD 1.1 till 1.5 are all trained on a batch size of 2048. The version pretty much everyone uses (1.5) is first pretrained at a resolution of 256x256 for 237K steps on laion2B-en, at the end of those training steps it will have seen roughly 500M images in laion2B-en. After that it is pre-trained for 194K steps on laion-high-resolution at a resolution of 512x512, which is a subset of 170M images from laion5B. Finally it is trained for 1.110K steps on LAION aesthetic v2 5+. This is easily verified by taking a glance at the model card of SD 1.5. Though that one doesn't specify for part of the training exactly which aesthetic set was used for part of the training, for that you have to look at the CompVis github repo. Thus at the end of it all both the most recent images and the majority of images will have come from LAION aesthetic v2 5+ (seeing every image approx 4 times). Realistically a lot of the weights obtained from pretraining on 2B will have been lost, and only provided a good starting point for the weights.

stable diffusion · CompVis/stable-diffusion@2ff270f · GitHub

Here for example --> https://github.com/CompVis/stable-diffusion