Oreilly – Inside Deep Learning, Video Edition 2022-6

Published on: 2024-10-12 19:36:17

Categories: 28

Description

Inside Deep Learning course, Video Edition. This course takes you on a journey through the world of modern deep learning theory and practice and helps you apply innovative techniques to solve everyday data problems. In this course, you will learn how to use PyTorch to implement deep learning, choose the right deep learning components, train and evaluate a deep learning model, tune deep learning models for maximum performance, deep learning terminology Understand and adapt existing Python code to solve new problems. About technology: Deep learning doesn’t have to be a black box! Knowing how your models and algorithms work gives you more control over the results. And you don’t need to be a mathematician or senior data scientist to understand what’s going on inside a deep learning system.

What you will learn:

Choosing the right components for deep learning
Training and evaluation of a deep learning model
Fine-tune deep learning models for maximum performance
Understanding deep learning terminology

This course is suitable for people who:

They want to understand the complex concepts of deep learning in simple language.
Looking for hands-on learning to implement deep learning with PyTorch.
Interested in solving real-world problems using deep learning.

Details of the course Inside Deep Learning, Video Edition

Publisher: Oreilly
Lecturer: Edward Raff
Training level: beginner to advanced
Training duration: 15 hours and 53 minutes

Course headings

Part 1. Foundational methods
Chapter 1. The mechanics of learning
Chapter 1. The world as tensors
Chapter 1. Automatic differentiation
Chapter 1. Optimizing parameters
Chapter 1. Loading dataset objects
Chapter 1. Summary
Chapter 2. Fully connected networks
Chapter 2. Building our first neural network
Chapter 2. Classification problems
Chapter 2. Better training code
Chapter 2. Training in batches
Chapter 2. Summary
Chapter 3. Convolutional neural networks
Chapter 3. What are convolutions?
Chapter 3. How convolutions benefit image processing
Chapter 3. Putting it into practice: Our first CNN
Chapter 3. Adding pooling to mitigate object movement
Chapter 3. Data augmentation
Chapter 3. Summary
Chapter 4. Recurrent neural networks
Chapter 4. RNNs in PyTorch
Chapter 4. Improving training time with packing
Chapter 4. More complex RNNs
Chapter 4. Summary
Chapter 5. Modern training techniques
Chapter 5. Learning rate schedules
Chapter 5. Making better use of gradients
Chapter 5. Hyperparameter optimization with Optuna
Chapter 5. Summary
Chapter 6. Common design building blocks
Chapter 6. Normalization layers: Magically better convergence
Chapter 6. Skip connections: A network design pattern
Chapter 6. 1 × 1 Convolutions: Sharing and reshaping information in channels
Chapter 6. Residual connections
Chapter 6. Long short-term memory RNNs
Chapter 6. Summary
Part 2. Building advanced networks
Chapter 7. Autoencoding and self-supervision
Chapter 7. Designing autoencoding neural networks
Chapter 7. Bigger autoencoders
Chapter 7. Denoising autoencoders
Chapter 7. Autoregressive models for time series and sequences
Chapter 7. Summary
Chapter 8. Object detection
Chapter 8. Transposed convolutions for expanding image size
Chapter 8. U-Net: Looking at fine and coarse details
Chapter 8. Object detection with bounding boxes
Chapter 8. Using the pretrained Faster R-CNN
Chapter 8. Summary
Chapter 9. Generative adversarial networks
Chapter 9. Mode collapse
Chapter 9. Wasserstein GAN: Mitigating mode collapse
Chapter 9. Convolutional GAN
Chapter 9. Conditional GAN
Chapter 9. Walking the latent space of GANs
Chapter 9. Ethics in deep learning
Chapter 9. Summary
Chapter 10. Attention mechanisms
Chapter 10. Adding some context
Chapter 10. Putting it all together: A complete attention mechanism with context
Chapter 10. Summary
Chapter 11. Sequence-to-sequence
Chapter 11. Machine translation and the data loader
Chapter 11. Inputs to Seq2Seq
Chapter 11. Seq2Seq with attention
Chapter 11. Summary
Chapter 12. Network design alternatives to RNNs
Chapter 12. Averaging embeddings over time
Chapter 12. Pooling over time and 1D CNNs
Chapter 12. Positional embeddings add sequence information to any model
Chapter 12. Transformers: Big models for big data
Chapter 12. Summary
Chapter 13. Transfer learning
Chapter 13. Transfer learning and training with CNNs
Chapter 13. Learning with fewer labels
Chapter 13. Pretraining with text
Chapter 13. Summary
Chapter 14. Advanced building blocks
Chapter 14. Improved residual blocks
Chapter 14. MixUp training reduces overfitting
Chapter 14. Summary
Appendix. Setting up Colab