sherlock113/bentosvd

This document demonstrates how to build an generative video application using BentoML, powered by diffusers and Stable Video Diffusion (SVD).

Prerequisites

• You have installed Python 3.9+ and pip. See the Python downloads page to learn more.
• You have a basic understanding of key concepts in BentoML, such as Services. We recommend you read Quickstart first.
• (Optional) We recommend you create a virtual environment for dependency isolation for this project. See Installation for details.
• If you want to test the service locally, a Nvidia GPU with 16G VRAM is required

Install dependencies

pip install -r requirements.txt

Run the BentoML Service

We have defined a BentoML Service in service.py. Run bentoml serve in your project directory to start the Service.

$ bentoml serve .

2024-01-19T07:29:04+0800 [INFO] [cli] Starting production HTTP BentoServer from "service:SVDService" listening on http://localhost:3000 (Press CTRL+C to quit)
Loading pipeline components...: 100%

The server is now active at http://0.0.0.0:3000. You can interact with it using Swagger UI or in other different ways.

CURL

curl -X 'POST' \
  'http://a4000box.jkdf.win:3000/generate' \
  -H 'accept: */*' \
  -H 'Content-Type: multipart/form-data' \
  -F 'image=@assets/sample.png;type=image/png' \
  -F 'decode_chunk_size=2' \
  -F 'seed=null' \
  -o generated.mp4

Deploy the application to BentoCloud

After the Service is ready, you can deploy the application to BentoCloud for better management and scalability. A configuration YAML file (bentofile.yaml) is used to define the build options for your application. It is used for packaging your application into a Bento. See Bento build options to learn more.

Make sure you have logged in to BentoCloud, then run the following command in your project directory to deploy the application to BentoCloud. Under the hood, this commands automatically builds a Bento, push the Bento to BentoCloud, and deploy it on BentoCloud.

bentoml deploy .

Note: Alternatively, you can manually build the Bento, containerize the Bento as a Docker image, and deploy it in any Docker-compatible environment. See Docker deployment for details.

Once the application is up and running on BentoCloud, you can access it via the exposed URL.