Package Details: stable-diffusion-intel 20220908-2

Git Clone URL: https://aur.archlinux.org/stable-diffusion-intel.git (read-only, click to copy)
Package Base: stable-diffusion-intel
Description: Image generator that uses stable diffusion, vino (Intel CPU) and includes "txt2img"
Upstream URL: https://github.com/bes-dev/stable_diffusion.openvino
Keywords: cpu diffusion stable
Licenses: Apache2
Submitter: xyproto
Maintainer: None
Last Packager: xyproto
Votes: 3
Popularity: 0.155246
First Submitted: 2022-09-06 11:20 (UTC)
Last Updated: 2022-12-12 14:56 (UTC)

Latest Comments

binarydepth commented on 2022-12-16 08:48 (UTC)

I am new to this. How do I change the model to a bigger one?, Since I have 32 GB of RAM. TY for this package.

bazza commented on 2022-12-15 17:12 (UTC)

It doesn't work with python 3.10 but with 3.9

martin3dimitrov commented on 2022-12-10 21:45 (UTC)

Seems like the latest python-diffusers breaks txt2img I get the following error:

None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
Traceback (most recent call last):
  File "/opt/stable-diffusion-intel/txt2img", line 74, in <module>
    main(args)
  File "/opt/stable-diffusion-intel/txt2img", line 17, in main
    scheduler = LMSDiscreteScheduler(
  File "/usr/lib/python3.10/site-packages/diffusers/utils/dummy_torch_and_scipy_objects.py", line 11, in __init__
    requires_backends(self, ["torch", "scipy"])
  File "/usr/lib/python3.10/site-packages/diffusers/utils/import_utils.py", line 281, in requires_backends
    raise ImportError("".join(failed))
ImportError: 
LMSDiscreteScheduler requires the PyTorch library but it was not found in your environment. Checkout the instructions on the
installation page: https://pytorch.org/get-started/locally/ and follow the ones that match your environment.


real    0m0.922s
user    0m0.827s
sys 0m0.086s

I tried to add pytorch as well but did not fix. Downgrading to python-diffusers 0.2.4 from the old PKGBUILD seems to solve this

xyproto commented on 2022-09-21 12:42 (UTC)

@WFCody I know! A lot of fun!

"img2img" is not included in the stable_diffusion.openvino repository, but "demo.py" (which is packaged as txt2img) takes an "--init-image" flag that might, just maybe, serve the same purpose.

WFCody commented on 2022-09-20 07:15 (UTC)

This is so much fun! Is it possible to run img2img with this implementation too?

Also, I had to edit the txt2img script in usr/bin because the += was not parsed correctly (perhaps because I use dash-static-musl as /bin/sh)