pixel-art-diffusion-api
How to run
- Install conda
- Run conda env create -f environment.yml (windows_environment.yml if on windows)
- python server.py
Scripts examplenation
- disco_xform_utils.py: is generated by the api.py
- main.py: test script to run the api
- pixel_art_diffusion.py: runs the generation/installation functions
- server.py: is the main server/api handler, it is using fastapi and uvicorn
- routes:
- /prompt_tags -> returns the available tags for the prompts (#bg, #pixelart etc)
- /generate (s: str (prompt), steps: int, seed: int) -> adds the generation task in the queue, returns the task id (data.id)
- /generate_result (query_id: str) -> returns the current state of task, if it is still running or the result has been taken, then it returns the status (running/finished), otherwise it returns the blob/byte array of the generated file