Deep Learning Practice lab

There are 6 labs in the course.

Lab1: BackPropagation

Implement forward path, backward propagation, and weight update to NN by using numpy.
- Architecture: linear layer
- Activation function: sigmoid & ReLU
- Loss function: Cross Entropy & MSE

Lab2: EEG Classification

Learn how to use pytorch to implement EEGNet, DeepConvNet, and custom dataset.

Lab3: Diabetic Retinopathy Detection (classification)

Learn how to implement ResNet and do classification to Diabetic Retinopathy task.

Lab4: Conditional Sequence to Sequence VAE

Implement data preprocessing in Character-Level
Implement seq2seq model(LSTM) combine w/ conditional VAE

Lab5: Conditional GANs

Reference to pytorch GAN to implement ACGAN & WGAN-GP
Customize multi-label dataset

Lab6: Play LunarLander by DQN, DDQN, DDPG

Learn how to implement DQN, DDQN, DDPG to play LunarLanderV2