TensorFlow.js Deployment Guide

Upload a Trained Model

1. Fork this repository
2. Create a new file mobilenet_1/README.md
3. Extract model.zip
4. Upload all files into mobilenet_1 folder
5. Click model.json > Raw
6. Copy raw URL of model.json
7. Download and extract this forked repository
8. Replace the model URL in index.js at tf.loadModel()

Setup Environment

1. Install Docker

• Ubuntu: https://docs.docker.com/install/linux/docker-ce/ubuntu/
• macOS: https://docs.docker.com/docker-for-mac/install/
• Microsoft Windows: https://docs.docker.com/docker-for-windows/install/

2. Pull Docker image (~200MB)

docker pull mrteera/tfjs:base

3. Run Docker

docker run -it --rm -v <tfjs-pacman-full-path>:/root/tfjs-pacman -p 1234:1234 mrteera/tfjs:base /bin/bash

For Windows

docker run -it --rm -v 'C:\ Downloads\tfjs-pacman:/root/tfjs-pacman' -p 1234:1234 mrteera/tfjs:base /bin/bash

For Windows (Legacy Docker)

docker run -it --rm -v '/c/ Downloads/tfjs-pacman:/root/tfjs-pacman/my_code' -p 1234:1234 mrteera/tfjs:base /bin/bash

4. install TensorFlow.js

cd /root/tfjs-pacman
cat index.js | grep dense_3   # only for windows legacy
yarn add @tensorflow/tfjs
yarn add @tensorflow/tfjs-node
yarn

5. Run yarn watch