Hanusz Leszek

Results 85 comments of Hanusz Leszek

Some tips for the CI: - always run `make check` before a commit to ensure that the linting issues get corrected. - The first time, it is a good idea...

Apparently [it's working with diffusers now](https://github.com/huggingface/diffusers#new--stable-diffusion-is-now-fully-compatible-with-diffusers). We just have to wait for the weights.

In the meantime, as a workaround, I simply [added an access code](https://github.com/leszekhanusz/diffusion-ui-backend/releases/tag/v0.2.0) in the gradio inputs.

̶@̶C̶i̶t̶o̶ ̶a̶n̶y̶ ̶i̶d̶e̶a̶ ̶w̶h̶y̶ ̶p̶y̶p̶y̶3̶ ̶c̶a̶n̶n̶o̶t̶ ̶f̶i̶n̶d̶ ̶g̶r̶a̶p̶h̶q̶l̶-̶c̶o̶r̶e̶ ̶v̶e̶r̶s̶i̶o̶n̶ ̶3̶.̶3̶.̶0̶a̶1̶ ̶?̶ Nevermind, it's because pypy3 still uses Python 3.6

> @leszekhanusz is this PR ready to ship? I'd love to get a new release of gql with a setup.py version bump of graphql-core that allows 3.3 to work (https://github.com/graphql-python/gql/blob/master/setup.py#L6)....

> A pre-release of gql would be awesome if it makes sense +1 [Pre-Release v3.5.0a0](https://github.com/graphql-python/gql/releases/tag/v3.5.0a0) is now available.

I just noticed that in PR #549, the code was not duplicated exactly the same way. Sometimes multiplying by `batch_size`, sometimes not. Proof again than duplicating code is never good...

I've read your blog post and you make some good points but some of it seems to contradict you: > we decided to adopt a different design principle which we...

@okalldal [it's possible to merge all the existing pipelines into a single one](https://github.com/leszekhanusz/diffusion-ui-backend/blob/eebb607244c183efb0eab2b5f3c367549ab0f56b/src/diffusionui/pipelines/pipeline_stable_diffusion.py#L26). Available [here](https://github.com/leszekhanusz/diffusion-ui-backend/tree/main/src/diffusionui/pipelines).

For img2img and inpainting you have to provide a generator instead of latents. You can [take a look on how I did](https://github.com/leszekhanusz/diffusion-ui-backend/blob/c980af79fc83d75bb08de62209238be9e6a1f490/src/diffusionui/gradio_interface.py#L12) it in diffusion-ui-backend. In my case you either...