Hands-on introduction to Deep Learning with Keras and Tensorflow
summary
Hands-on notebooks show building models with Keras with Tesorflow backend, training dense and convolutional nets, and using batch normalization and dropout; MNIST-style digits reach ~90% while a harder dataset stalls near 30% but MobileNet transfer learning boosts results toward ~80%. Lesson: compile/optimizer choices (Adam), batch size and epochs change stability, 5,000 labeled examples is a rough start point for supervised tasks and millions are needed to match human performance.
AI-written summary · about
chapters (12)
- 0:00Deep learning with Keras/TensorFlow
- 1:15Deep learning breakthroughs
- 10:35Backpropagation training
- 20:08Deep learning vs other methods
- 23:29Deep learning tools overview
- 30:07Keras Sequential API
- 39:54Model layers and parameters
- 50:12Training and learning curves
- 56:40Batch norm and dropout
- 1:05:59CNNs for image data
- 1:15:05Transfer learning models
- 1:26:28Summary
related talks
01h 04mWhat you always wanted to know about Deep Learning, but were afraid to ask
35mDeconstructing Deep Learning
44mTensorflow, Tensorflow, What is it that you see?
01h 06mDeep Learning with TensorFlow
02h 41mHello TensorFlow: 3 Workshops to Get Started with TensorFlow 2.0
01h 37mTen Steps to Keras
49mDeep Learning With Python & Tensorflow
45mThe most soft and fluffy intro into Machine Learning and Deep Neural Networks