Jamuin-ML

Introduction

Team ID:C23-PS445

Contributors

Name	Specialization	Role	Profile
Yohanes Egi Pratama Yudoutomo	Machine Learning	ML Engineer	Github
Dian Alhusari	Machine Learning	ML Engineer	Github
Muhammad Faishal Azhar Suherman	Cloud Computing	Cloud Engineer (ML Deployment)	Github

Model Description:

a machine learning model was created to classify images of herbal ingredients that we often encounter to be recommended into herbal products, including:

Built with

• Python
• TensorFlow
• Flask
• Google Cloud Run
• Google Colab

Dataset:

dataset of herbal ingredients obtained by scraping, and manually data

Preview of the image and data used are shown in the picture below:

Kunyit

Arsitektur Model

Arsitektur Model

Make RESTfull API with Flask and Cloud Run

1. Prepare prediction model in "h.5" format, file are stored in "ML-Backend" folder
2. Write main.py base on machine learning testing model, files are saved in the "ML-Backend" folder
3. Create file named "requirement.txt" for library you need for running our code
4. Create file named "Dockerfile" for run system in our container
5. Create file named ".dockerignore" for ignore system to ignore spesific file.
6. Create folder static/uploads to save photos for prediction progress.
7. Create new Project in Google Cloud Platform
8. Active Cloud Run API and Cloud Build API
9. Install and init Google Cloud SDK (Use this link : https://cloud.google.com/sdk/docs/install)
10. Use Cloud Build to import our code to our cloud services ( gcloud builds submit --tag gcr.io/<project_id>/<function_name>)
11. Use Cloud Run to deploy our API ( gcloud run deploy --image gcr.io/<project_id>/<function_name> --platform managed )