Torch Deep Learning Framework Video Tutorials

Learn deep learning fundamentals through this comprehensive video tutorial series covering Torch, the Lua-based scientific computing framework. Master Lua programming basics before diving into neural network concepts including forward and backward propagation using the nn package. Explore convolutional neural networks (CNNs) from basic principles through internal mechanisms, model architectures, training procedures, and loss functions. Progress to recurrent neural networks (RNNs) covering vector and sequence processing, forward and backward propagation mechanics, implementation using the nngraph package, and training methodologies. Work with the image package for computer vision tasks and gain hands-on experience with practical implementations throughout the 8 hour 27 minute tutorial series developed by Alfredo Canziani.

Syllabus

Practical 1.0 – Lua
Practical 1.1 – Torch
Practical 1.2 – image package
Practical 2.0 – NN forward
Practical 2.1 – NN backward
Practical 2.2 – nn package
Practical 3.0 – CNN basics
Practical 3.1 – CNN internals
Practical 3.2 – CNN models
Practical 3.3 – CNN training
Practical 3.4 – CNN loss
Practical 4.0 – RNN, vectors and sequences
Practical 4.1 – RNN forward and backward
Practical 4.2 – RNN nngraph package
Practical 4.3 – RNN training