Lec 12 | Sequence-to-Sequence Models

This lecture covers sequence-to-sequence modeling, explaining the encoder-decoder architecture, training, and decoding strategies like greedy search, exhaustive search, and beam search. It highlights the trade-offs between efficiency and accuracy in sequence generation and explores techniques to improve decoding quality.

🎓 Lecturer: Tanmoy Chakraborty [https://tanmoychak.com]
🔗 Get the Book: https://tanmoychak.com/llmbook
📚 Suggested Readings:
- Sequence to Sequence Learning with Neural Networks [https://arxiv.org/pdf/1409.3215]
- Neural Machine Translation by Jointly Learning to Align and Translate [https://arxiv.org/pdf/1409.0473]
- Chapter-5, Intro to LLM, Section 5.3 (Sequence-to-Sequence Models) [https://tanmoychak.com/llmbook]

This lecture is about sequence-to-sequence modeling and the encoder-decoder framework in this lecture. Dive into training methods and decoding strategies like greedy search, exhaustive search, and beam search. Understand the balance between efficiency and accuracy in sequence generation and discover techniques to enhance decoding performance.

Видео Lec 12 | Sequence-to-Sequence Models канала NPTEL IIT Delhi