Skip to content

Latest commit

 

History

History
93 lines (78 loc) · 3.27 KB

README.md

File metadata and controls

93 lines (78 loc) · 3.27 KB

dockerized dev environment / playground for dalle-mini

flask and react web app to generate images from text


app


SOME EXAMPLE IMAGES GENERATED WITH THIS REPO


clown
"scary clown playing a cello painted in the style of van gogh"

frog
"frog riding a unicorn standing on top of a mountain of cookies"

seal_otter
"otter and seal sitting on an iceberg drawn in cartoon style"

super_spider
"superman punching spiderman on an iceberg"

CLONE


  • git clone https://github.com/revzim/dalle-mini-server.git

RUN


  • DOCKER
    • docker compose
      • start
        • docker compose up
      • rebuild all
        • docker compose build
      • rebuild server
        • docker compose build dalle_server_flask
      • rebuild client
        • docker compose build dalle_client_react
      • to force rebuild of entire container add --no-cache
    • docker only
      • server/ & client/ both contain respective Dockerfiles
    • dalle-mini api endpoints local
    • dalle-client frontend react app local
  • LOCAL (LINUX/UNIX*)
    • server
      • pytorch required
      • cd server
      • python venv recommended python -m venv .
      • pip install -r requirements.txt
      • python app.py --port=8080
    • client
      • cd client
      • yarn or npm install
      • yarn start or npm run start

-important installation/run note-

  • dalle-mini uses large models that sometimes can take a couple minutes to load, please be patient while the server starts

for testing / educational purposes only


dalle-mini trained models are used to generate images from a text prompt

dev build pack inspired by dalle-playground


author: revzim