Nathan Lambert

Great. I want to take a look, and I have chatted with @danijar who didn't know it was being worked on. Let's see if I can un-stick it and if...

@mohakbhardwaj -- I haven't made the time to make the PR. Happy to provide feedback if you take a stab at it?

Yeah @anton-l and @Suhail I agree with both points. I'll do some experiments and share the results + tests

Yeah I'll come back to this when I have time. Mostly this was a learning side-project for me, so leaving open is preferred.

What do people think. Is it worth spending time fleshing this out? SDE VE is the improved version of this scheduler, so unless we want to include it for completeness,...

Hi @zcswdt -- you need to get the model directly from the hub (https://huggingface.co/runwayml/stable-diffusion-v1-5). This involves making an account, accepting the conditions / license, and then downloading directly or using...

This is a great start. I noticed a small bug and created a [colab](https://colab.research.google.com/drive/1M1vdP6mFBxVTLbIHrNhWtfLTKO4jaegy?usp=sharing) to play with the sampling notebook. The core code from algorithm 2 should be updated like...

@patrickvonplaten I looked really closely at this and I don't see a reason that dance_diffusion test is failing. Do you agree? Hmmmm... testing is really tricky here.

Two todo's in order to merge this: 1. update the assertion test in dance diffusion, as it's pretty clear the underlying model logic is the same (multiple folks have checked...

@patrickvonplaten FYI I checked that the harmoni ai slow test passes. I changed the values of the fast dance_diffusion_pipeline test, which I think is needed because pytorch generates values differently...