---
title: Week 3
---


Oct 09
: **Lecture**{: .label .label-green }[Throughput, Latency, and Efficient Inference](#)
  - What are the different types of inference workloads? Understanding throughput and latency.
  - Reviewing Autoregressive Generation
  - Inference Efficiency Properties of RNNs vs. Transformers
  - Techniques for Efficient Inference: KV-caching, FlexGen, vLLM, etc.


Oct 11
: **Lecture**{: .label .label-green }[Alternate Architectures to Transformers & Course Project Intro](#)
  - Reviewing the Efficiency Properties of Convolutions vs. RNNs vs. Transformers
  - Recent Subquadratic Models
  - Hardware Aware Algorithms for the Subquadratic Models
  - Course Project Introduction