psychedelicious

@keturn the canvas relies on lossless images. Do you have any ideas for working around that and also reducing image size?

Yes, but also - the canvas consists of images layered on top of one another. Compression can lead to artifacts on the edges of images and visible seams. I'm generally...

I've downloaded the trinart vae and cannot get it to show up in the model manager at all - cannot reproduce this. Can you please post your `configs/models.yaml` file?

@lstein I'm not sure how to approach this. The offending `models.yaml` is provided in a comment above

This is expected to be resolved as the fix was in diffusers v13.

This is a known upstream issue with diffusers. There is a fix which we expect to be in the next diffusers release.

FWIW, the UI will be using a `husky` to format and lint via pre-commit hook. ~~It's being used in my [nodes migration](https://github.com/invoke-ai/InvokeAI/pull/2451) branch for now.~~ This is now in main.

@realstevewarner @lstein I'm pretty sure that the metadata *is* readable, but you have to refresh the page. The server doesn't read the metadata correctly when first uploading an image. Going...

@realstevewarner We have v2.3.0 pending release. It will not be fixed in that. The bug is in the current web server - we are really trying to not making any...

@realstevewarner You shouldn't lose anything on refresh - that's a different bug. Please try `Reset Web UI` from `Settings` (gear icon @ top right) then refresh the page (this will...