CONTINUUM Books
Efficient Transformer Inference
Lecture notes
Lecture notes
Efficient Transformer Inference
Lecture notes
Published 12 August 2026 · 142 pages · English
ISBN 978-0-00-100004-3 · DOI 10.xxxx/ctm.bk.2026.0004
Sample book. Title, authors and content are fictional demonstration material.
About this book
Lecture notes from a graduate course on making transformer inference fast and small: attention variants, caching, speculative decoding, quantisation, batching and scheduling. Includes problem sets and a reading list.
Keywords inference · transformers · speculative decoding · quantisation
Read or download
efficient-transformer-inference-notes.pdf
efficient-transformer-inference-notes.html
Contents
- 1Where the time goespp. 1–20PDF
- 2Attention and KV cachingpp. 21–52PDF
- 3Quantisationpp. 53–84PDF
- 4Speculative decodingpp. 85–110PDF
- 5Serving and schedulingpp. 111–142PDF
About the author
- Rahul Desai · Meridian University, India
Works on decoding algorithms and systems optimisation for transformer inference.
How to cite
Desai, R. (2026). Efficient Transformer Inference: Lecture notes (Version 1.1). CONTINUUM Books. doi:10.xxxx/ctm.bk.2026.0004