Will Berman

SVD operator

13

### Semantics: The SVD operator covers pytorch, numpy, and tensorflow’s SVD semantics. Numpy and tensorflow use the same `compute_uv` flag for computing just the singular values. Pytorch uses two different...

[WIP] VQ-diffusion

8

NOTE: updating PR description and moving verification code to [notebook](https://github.com/williamberman/vq-diffusion-notebook) ## Porting the VQ-diffusion VQVAE for the ITHQ dataset to diffusers. Add `convert_vq_diffusion_to_diffusers.py` script: This script initially only converts the...

[wip] kakaobrain unCLIP

4

##### Scoping - [x] Scope prior Transformer - [x] Scope decoder Unet - [x] Scope super resolution 64->256 Unet - [x] Scope super resolution 256 ->1024 Unet - [x] Scope...

[testing] tiny stable diffusion pipe throws error when used with safety checker

2

### Describe the bug The tiny stable diffusion pipeline we use for some tests throws an error when used with the safety checker. It currently doesn't throw an error in...

The module with `PipelineTesterMixin` already sets this. We shouldn't repeatedly set this in all tests and instead consolidate where we set it

Marginal improvements over xformers, note that xformers itself isn't a tremendous improvement over vanilla attention. We have relatively small sequence lengths of 256 for self attention and 256x77 for cross...

this will help for benchmarking since we don't have to include xformers in it :P Testing the MOVQ: ``` from muse.modeling_movq import MOVQ from PIL import Image import numpy as...