advanced-deep-learning

Advanced Deep Learning lecture from Friedrich-Alexander University Erlangen Nuernberg. These exercises done by Abdallah Eid and Ahmed Sheta.

Exercise 1

Exercise one include:

• Vision Transformer (ViT) for Self-attention for image classification.
• Cross-Attention Multi-Scale ViT: Adding cross-attention

Exercise 2

Exercise two include Diffusion Models for image generation.

Exercise 3

Exercise three include:

• Energy-based Models (EBM)
• Sampling the model with Stochastic Gradient Langevin Dynamics (SGLD)
• Joint energy-based model (JEM)
• Out-of-distribution (OOD) detection with EBMs