ML_Final-Project

This guide is for Windows. For Linux and MacOS, it should be easier :)

Installation

1. Install Anaconda

2. Install Postman

3. Install Visual Studio Community (Core Editor Only)

4. Install CUDA Toolkit 11.0 Update

5. Download cuDNN 8.0.5 for CUDA 11.0

   • Extract cuDNN zip file, copy all folders in cuda folder to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0 (select replace)

   • Add these to Path environment variables

     C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\bin
     C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\libnvvp
     C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\extras\CUPTI\lib64
     C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\include

   • Reboot

6. Create virtual environment

   • conda create --name tf2.4 python==3.8
   • conda activate tf2.4
   • pip install tensorflow==2.4.0
   • pip install opencv-python
   • pip install numpy

7. Install Jupyter Notebook

   • conda install -y jupyter
   • conda install -y nb_conda

8. Open the project

   cd ML_Final-Project
   jupyter notebook Final_Project_Bangkit.ipynb

Testing

• Convert image to base64 format using deploy/tobase64.py by changing PATH variable to your image file location. Then run this command:

  cd deploy
  python tobase64.py

• Generate JSON

  python generateJSON.py

• run the flask app on deploy/deploy.py

  python deploy.py

• Copy everything from deploy/data.json to body like this

  

• Test to send POST method to <url-where-the-flask-app-running>/predict on Postman

Deployment

We use Google Cloud Run for the deployment of this project

• requirements.txt
• cloudmigrate.yaml
• Dockerfile