Hamish Friedlander

@tonetechnician @hlky's fork is pretty old now. My repo is based on Diffusers 0.4.2, so upgrading shouldn't be too painful. However I've backported 0.3.0-style scheduler support into my repo to...

If you're using my update of the schedulers, but not my unified_pipeline, it probably won't work. (That's my guess for why camo-dog).There's quite a few special handling cases in the...

@patrickvonplaten tricky to work around that constraint You can see here: https://github.com/hafriedlander/stable-diffusion-grpcserver/blob/776502981a872ddaa2f6b7c2ec9a58adf0154372/sdgrpcserver/pipeline/old_schedulers/scheduling_dpm2_discrete.py#L168 The scheduler calculates a predicted state, and then re-calculates the noise prediction for the next step based on...

This (internal? unclear) API change will have also broken all custom schedulers, since most of them rely on set_format to convert from numpy to tensor. Might need a stronger warning?...

Actually, I'll suggest a different fix. Put the old behaviour back into scheduling_utils - you don't use it in the core schedulers at all, so it won't matter, and it'll...

Hi. Sorry for delay on replay. I am using the additional schedulers Hlky put together at https://github.com/huggingface/diffusers/compare/main...hlky:diffusers:main However I consider this solved now for two reasons: - I just copied...

https://github.com/hafriedlander/diffusers/blob/stable_diffusion_2/scripts/convert_original_stable_diffusion_to_diffusers.py Notes: - Only tested on the two txt2img models, not inpaint / depth2img / upscaling - You will need to change your text embedding to use the penultimate layer...

Here's an example of accessing the penultimate text embedding layer https://github.com/hafriedlander/stable-diffusion-grpcserver/blob/b34bb27cf30940f6a6a41f4b77c5b77bea11fd76/sdgrpcserver/pipeline/text_embedding/basic_text_embedding.py#L33

You need to use the absolute latest Diffusers and merge this PR (or use my branch which has it in it) https://github.com/huggingface/diffusers/pull/1386

(My branch is at https://github.com/hafriedlander/diffusers/tree/stable_diffusion_2)