#11 Live: Training Models in TensorFlow.js, Web Coding 101 & Loading ResNet50 from TensorFlow Hub
🤖 Let's write a custom linear regression model (and use a pre-trained ResNet network) in a web browser with TensorFlow.js. Watch this to get an overview of using HTML & JavaScript to write front-end code and create web-based interactive apps.
🤖 Code https://github.com/nonoesp/live/tree/...
💬 Join the conversation on Discord / discord
🧠 Machine Learning Series: • 🧠 Machine Intelligence
🔴 Live Streams: • 🧠 Live Streams
🍃 Getting Simple: https://gettingsimple.com
🎙 Podcast: https://gettingsimple.com/podcast
🗣 Ask Questions: https://gettingsimple.com/ask
💬 Discord: / discord
👨🏻🎨 Sketches: https://sketch.nono.ma
✍🏻 Blog: https://nono.ma
🐦 Twitter: / nonoesp
📸 Instagram: / nonoesp
0:00 Start
3:03 Intro
4:03 PEmbroider
5:42 Let's code
7:03 PEmbroider on GitHub
11:52 Live 11
14:08 TensorFlow Hub
18:04 Glitch
21:26 Website Development 101 - HTML JavaScript
24:58 Adding TensorFlow.js to our Website
27:01 Inspecting Browser Network Calls
34:02 Using the Browser JavaScript Console
38:44 Node.js REPL
39:40 Python REPL
40:50 TensorFlow.js in the Console
50:04 Linear Regression in TensorFlow.js
53:25 Define Your Model
1:00:28 Define Your Training Data
1:01:01 Train Your Model
1:07:28 Add a Callback After Each Training Epoch
1:09:57 Training Callbacks
1:15:34 Understanding our Linear Input Data
1:18:11 Training Results
1:27:04 Request Predictions from Your Trained Model
1:41:07 Adding More Data to Converge Faster
1:44:06 Loading ResNet50 from TensorFlow Hub with TF.js
2:01:01 Pushing Today's Code to GitHub
2:10:15 Next Week
2:13:22 Outro