stable_diffusion.openvino

Implementation of Text-To-Image generation using Stable Diffusion on Intel CPU or GPU.

Requirements

• Linux, Windows, MacOS
• Python <= 3.9.0
• CPU or GPU compatible with OpenVINO.

Install requirements

• Set up and update PIP to the highest version
• Install OpenVINO™ Development Tools 2022.3.0 release with PyPI
• Download requirements

python -m pip install --upgrade pip
pip install openvino-dev[onnx,pytorch]==2022.3.0
pip install -r requirements.txt

Generate image from text description

usage: demo.py [-h] [--model MODEL] [--device DEVICE] [--seed SEED] [--beta-start BETA_START] [--beta-end BETA_END] [--beta-schedule BETA_SCHEDULE]
               [--num-inference-steps NUM_INFERENCE_STEPS] [--guidance-scale GUIDANCE_SCALE] [--eta ETA] [--tokenizer TOKENIZER] [--prompt PROMPT] [--params-from PARAMS_FROM]
               [--init-image INIT_IMAGE] [--strength STRENGTH] [--mask MASK] [--output OUTPUT]

optional arguments:
  -h, --help            show this help message and exit
  --model MODEL         model name
  --device DEVICE       inference device [CPU, GPU]
  --seed SEED           random seed for generating consistent images per prompt
  --beta-start BETA_START
                        LMSDiscreteScheduler::beta_start
  --beta-end BETA_END   LMSDiscreteScheduler::beta_end
  --beta-schedule BETA_SCHEDULE
                        LMSDiscreteScheduler::beta_schedule
  --num-inference-steps NUM_INFERENCE_STEPS
                        num inference steps
  --guidance-scale GUIDANCE_SCALE
                        guidance scale
  --eta ETA             eta
  --tokenizer TOKENIZER
                        tokenizer
  --prompt PROMPT       prompt
  --params-from PARAMS_FROM
                        Extract parameters from a previously generated image.
  --init-image INIT_IMAGE
                        path to initial image
  --strength STRENGTH   how strong the initial image should be noised [0.0, 1.0]
  --mask MASK           mask of the region to inpaint on the initial image
  --output OUTPUT       output image name

Examples

Example Text-To-Image

python demo.py --prompt "Street-art painting of Emilia Clarke in style of Banksy, photorealism"

Example Image-To-Image

python demo.py --prompt "Photo of Emilia Clarke with a bright red hair" --init-image ./data/input.png --strength 0.5

Example Inpainting

python demo.py --prompt "Photo of Emilia Clarke with a bright red hair" --init-image ./data/input.png --mask ./data/mask.png --strength 0.5

Performance

CPU	Time per iter	Total time
AMD Ryzen 7 4800H	4.8 s/it	2.58 min
AMD Ryzen Threadripper 1900X	5.34 s/it	2.58 min
Intel(R) Core(TM) i7-4790K @ 4.00GHz	10.1 s/it	5.39 min
Intel(R) Core(TM) i5-8279U	7.4 s/it	3.59 min
Intel(R) Core(TM) i5-8569U @ 2.8GHz (MBP13-2019)	6.17 s/it	3.23 min
Intel(R) Core(TM) i7-1165G7 @ 2.80GHz	7.4 s/it	3.59 min
Intel(R) Core(TM) i7-11800H @ 2.30GHz (16 threads)	2.9 s/it	1.54 min
Intel(R) Core(TM) i7-1280P @ 1.80GHz (6P/8E)	5.45 s/it	2.55 min
Intel(R) Xeon(R) Gold 6154 CPU @ 3.00GHz	1 s/it	33 s
Intel Arc A770M	6.64 it/s	7.53 s

Acknowledgements

• Original implementation of Stable Diffusion: https://github.com/CompVis/stable-diffusion
• diffusers library: https://github.com/huggingface/diffusers

Disclaimer

The authors are not responsible for the content generated using this project. Please, don't use this project to produce illegal, harmful, offensive etc. content.