Virtual Workshop 1.0

Overview

This virtual workshop walks you through the process of building and deploying a simple machine learning model using Teachable Machine and TensorFlow.js. We will create a model, export it, and host the final solution using GitHub and Netlify.

Steps

1. Create a Teachable Machine Model

• Sample Insert: Upload a set of sample images or videos for each class to train the model.
• Train Model: Once the samples are uploaded, click on the "Train Model" button to begin training.

2. Save and Export the Model

• Save the Model: Once the model is trained, save it to your Google Drive for future use.
• Export Model: Export the model for use in web applications. Select the option to export it as a TensorFlow.js model.

3. Upload Model to Cloud

Upload the model files (including model.json and weights) to a cloud storage service such as Google Drive, GitHub, or any other cloud platform that allows public file hosting.

4. Add the Code

• Get TensorFlow.js Code: Copy the generated TensorFlow.js code snippet from the Teachable Machine export.

• Create an index.html File: In your local project folder, create an index.html file. Paste the copied TensorFlow.js code into this file.

  You can add the following line to display predictions as a percentage:

  const classPrediction = prediction[i].className + ": " + (prediction[i].probability * 100).toFixed(2) + "%";

5. Upload to GitHub

• Create a new repository on GitHub.
• Push the index.html file along with the model files to this repository.

6. Deploy with Netlify

• Go to Netlify.
• Connect your GitHub repository to Netlify.
• Deploy the project by following Netlify's step-by-step deployment instructions.