This is a BentoML service that demonstrates how to parse image using multi-modal LLM and extract useful information from them.
The work is based on @PsiACE's blog post.
This project is managed by PDM, install it first.
Install dependencies:
pdm installCreate .env file with credentials:
cp .env.example .env
# Complete the OPENAI_API_KEY in the .env fileStart the development server:
pdm dev-
Go to BentoCloud and get an account.
-
Login to BentoCloud:
pdm run bentoml cloud login
-
Click "Secret" in the sidebar and create an "OpenAI" secret with your API key.
-
Deploy the service to BentoCloud:
pdm deploy --secret <your-secret-name>
This work is released under Unlicense. You can use it for any purpose without any restriction.