Building a machine learning app | 10 hour coding livestream

Опубликовано: 20 Январь 2025
на канале: Daniel Bourke

42,439

1.7k

Time to celebrate 100k subscribers! And what better way to do it than with a coding livestream?

Tune in for 10 hours straight (give or take a bit for dancing breaks/food etc).

Get the code - https://github.com/mrdbourke/food-not...
Try the app - https://www.foodnotfood.app

Get the free stuff/surprises here - https://www.mrdbourke.com/100k-subs/

Links:
Learn ML (beginner-friendly courses I teach) - https://www.mrdbourke.com/ml-courses/
ML courses/books I recommend - https://www.mrdbourke.com/ml-resources/
Read my novel Charlie Walks - https://www.charliewalks.com

Connect elsewhere:
Web - https://dbourke.link/web
Twitter - / mrdbourke
Twitch - / mrdbourke
ArXiv channel (past streams) - https://dbourke.link/archive-channel
Get email updates on my work - https://dbourke.link/newsletter

Timestamps:
0:00 - Intro & hello
3:53 - Stream rules (10 pushups, squats, kicks, punches every hour)
5:00 - What we're going to build (food not food app)
10:58 - First commit
21:52 - Collecting a dataset (food and not food images)
2:20:30 - Food break
2:33:59 - Back to coding
3:36:09 - Downloading images
3:55:25 - Small model building
4:47:59 - Making train and test datasets
4:56:18 - Building binary classification model
5:29:54 - First model training
5:54:34 - Filtering food and non-food images from ImageNet
7:52:39 - Machine learning model deployment with TensorFlow Lite and TensorFlow JS
9:16:44 - Testing predictions with TensorFlow Lite model
9:40:39 - First prediction working on deployed model!!!
10:04:50 - Model deployed and live to the internet!!!