Web Application for deploying neural net model on Render
- basic html interface
- cam in or image input
- server.py to interface with server
- js to receive and analyse image
- relay audio out
The repository can be used as a starting point to deploy any model on Render.
Guide for production deployment :
- Fork this repository or clone the app/ folder into your Github account.
- Upload your trained model file.
- Upload the trained model file created with learner.export (for example export.pkl) to a cloud service.
- Edit the file server.py inside the app directory and update the export_file_url variable with the URL copied above. In the same file, update the line "classes" with the classes of your model.
- Create a new Web Service on Render and use the repo you created above.
Local testing : To run the app server locally, run this command in your terminal:
python app/server.py serve
Changes can be tested locally by installing Docker and using the following command:
docker build -t "repo-name" . && docker run --rm -it -p 5000:5000 "repo-name"
Officially time to shift from Heroku to Render :
- Although being expedient, moving from AWS to Render eliminated a lot of infrastructure complexity.
- I would prefer GCP being a student but it compares nothing against AWS Services.