SLIDE 1
Transfer Learning with Convolutional Neural Networks
- Off the shelf top notch
Convolutional Neural Networks ---- Off the shelf top notch - - PowerPoint PPT Presentation
Convolutional Neural Networks ---- Off the shelf top notch - - PowerPoint PPT Presentation
Transfer Learning with Convolutional Neural Networks ---- Off the shelf top notch performances Convolutional Neural Networks A breakthough Convolutional Neural Networks VGG-16 example Layers of Convolutional filters Bottleneck
SLIDE 2
SLIDE 3
Convolutional Neural Networks VGG-16 example
- Layers of Convolutional
“filters”
- Bottleneck architecture
- Magical ?
Features extraction Classification
SLIDE 4
CNNs – Inner working
SLIDE 5
CNNs – Feature Extraction
SLIDE 6
Training CNNs
- Big Dataset of 100 thousands images –
usually millions
- Labelled data Takes time to build
- Try out many different Network
architecture
- Hyperparameter value :
– training method, rate – Weights initial values
- No feature engineering
SLIDE 7
Training CNNs – Live view
SLIDE 8
ImageNet Database
SLIDE 9
SLIDE 10
SLIDE 11
SLIDE 12
From: AN ANALYSIS OF DEEP NEURAL NETWORK MODELS FOR PRACTICAL APPLICATIONS Alfredo Canziani & Eugenio Culurciello & Adam Paszke
SLIDE 13
Trained with ImageNet
SLIDE 14
SLIDE 15
SLIDE 16
SLIDE 17
SLIDE 18
SLIDE 19
SLIDE 20
SLIDE 21
SLIDE 22
- Less likely to overtrain
SLIDE 23
Alternatives
- CNN architecture & hyperparameter
“automatic” tuning
– AutoML (Architecture) – driven by “AI” – IBM Watson Suite (Hyperpameters) – Microsoft Custom Vision Services
- Drawbacks