RESTful API Server For Managing Triton Inference Server

Preparations

pip install -r requirements.txt
python3 manage.py makemigrations
python3 manage.py migrate

Run API Server

python3 manage.py runserver

Usage

Create Models

1. Fill basic configurations Link: http://127.0.0.1:8000/models/

2. Add input and output configurations Link:

   • http://127.0.0.1:8000/models/{model_name}/inputs/
   • http://127.0.0.1:8000/models/{model_name}/outputs/

3. Create new version and add model file Link: http://127.0.0.1:8000/models/{model_name}/versions/

4. (optional) Add custom files to each version Link: http://127.0.0.1:8000/models/{model_name}/versions/{version_id}/custom-files/

Create Repositories

Fill name and select models Link: http://127.0.0.1:8000/repositories/

Manage Triton Server

Execute managing operations Link: http://127.0.0.1:8000/operations/

Possible operations:

• create: create directory and all triton models for selected repository
• remove: remove all contents of selected repository
• start: start triton server using selected repository
• stop: stop current triton server