11. Sequence Models
RNN
- Process sequences
- Hidden state update
- Computational graph
- Truncated back-propagation through time
Sequence-to-sequence modeling
- Image captioning
- VQA
- VLN
LSTM
- Vanilla RNN vs. LSTM
- Gates explanation
- Gradient flow
GRU
Transformer
- Attention
- Transformer
- ViT & Swin