You Can Do Deep Learning!
summary
Demonstrates how to build a cats vs dogs image classifier end‑to‑end using Python tools: fastai on top of PyTorch, run in Kaggle kernels with a GPU enabled. Uses pretrained ResNet architecture and simple data wrappers to load labeled images, trains for two epochs in a few minutes and reports ~99% accuracy in the demo. Shows where to change learning rate, try other datasets or architectures, and points to community resources and reproducible kernels.
AI-written summary · about
chapters (17)
- 0:00Start
- 1:44Other applications
- 5:57$ gem install rails
- 7:03Magic!
- 8:15Common roadblocks
- 10:32the fast.ai formula
- 12:07GOOGLE CLOUD BIG DATA AND MACHINE LEARNING BLOG
- 14:10Keras Documentation
- 20:58Kaggle Kernels
- 22:54fast.ai library
- 24:17Communities
- 25:26Collection of Interactive Machine Learning Examples
- 26:18Near term questions
- 27:30Understand, implement, and explain
- 28:44Win an Emmy?
- 29:31Tackle your business problems
- 30:30Take on the giants
related talks
01h 00mA gentle introduction to Deep Learning
02h 26mPractical Deep Learning for Data Scientists
01h 01mA practical guide to deep learning
01h 01mA practical guide to deep learning
18mPutting the Fast in fast.ai - Machine Learning Models in 15 mins
29mComputer Vision Dog Breed Classification with Convolutional Neural Networks, TensorFlow and Kubeflow
47mDeep Learning with PyTorch
01h 26mDeep Neural Networks with PyTorch