kaito
kaito copied to clipboard
feat: Stable Diffusion
Currently this just supports text-to-image. Not Image Inpainting, Super-Resolution or Depth-to-Image
Need these dependencies on machine:
RUN apt-get update && apt-get install ffmpeg libsm6 libxext6 -y
Accelerate launch:
CUDA_VISIBLE_DEVICES="0" accelerate launch --num_processes 1 --num_machines 1 --machine_rank 0 inference_api.py --pretrained_model_name_or_path <MODEL_NAME>
Curl command
curl -X POST "http://localhost:5000/generate?prompt="microsoft%20employee%20riding%20a%20horse%20on%20mars"" > output.png
Example:
CUDA_VISIBLE_DEVICES="0" accelerate launch --num_processes 1 --num_machines 1 --machine_rank 0 an_inference.py --pretrained_model_name_or_path stabilityai/stable-diffusion-xl-base-1.0 --allow_remote_files --torch_dtype=float16 --use_safetensors --variant="fp16"
API Requests:
curl -X POST "http://localhost:5000/generate?prompt=astronaut%20riding%20a%20horse%20on%20mars%20eating%20a%20pizza" > output.png
curl -X POST "http://localhost:5000/generate?prompt=Kubernetes%20cluster%20orchestrating%20a%20dance%20party%20for%20AI-powered%20robots%20on%20a%20cloud" > output.png
curl -X POST "http://localhost:5000/generate?prompt=Microsoft%20engineer%20enjoying%20a%20lively%20celebration%20with%20their%20team%20at%20KubeCon,%20animated%20cartoon" > output.png
Codecov Report
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 65.33%. Comparing base (
de2b45c) to head (5d64ade). Report is 45 commits behind head on main.
Additional details and impacted files
@@ Coverage Diff @@
## main #282 +/- ##
==========================================
+ Coverage 65.27% 65.33% +0.06%
==========================================
Files 14 14
Lines 1123 1125 +2
==========================================
+ Hits 733 735 +2
Misses 352 352
Partials 38 38
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.