Media Summary: Speaker: Yu Cheng, Principal Researcher, Microsoft An agent written RMSNorm kernel hit 1.88x speedups on H100s. A finetuned Qwen3 0.6B hit 35% on LiveCodeBench. Neither ... This video reviews and discusses the paper Reformer: The Efficient
Research Talk Transformer Efficiency From - Detailed Analysis & Overview
Speaker: Yu Cheng, Principal Researcher, Microsoft An agent written RMSNorm kernel hit 1.88x speedups on H100s. A finetuned Qwen3 0.6B hit 35% on LiveCodeBench. Neither ... This video reviews and discusses the paper Reformer: The Efficient Thanks to ML6 for virtually hosting us tonight! For those who would like to attend live, have a look at ... Okay we're going to look at calculating the Reformer by Kitaev et al. (2020) tackled the inefficiencies of vanilla
... question taken from April today 2022 paper begin a 250 kingle face 50
![[REFAI Seminar 06/08/21] Transformer efficiency: From model compression to training acceleration](https://i.ytimg.com/vi/kdBEYwhG6mc/mqdefault.jpg)
![[T@W intro] Auke Wiggers — Efficient Transformers](https://i.ytimg.com/vi/s_yPF7nIJ7U/mqdefault.jpg)
![[Full Talk] Auke Wiggers (Qualcomm) — Transformers at Work](https://i.ytimg.com/vi/Z6XX2aNtg-g/mqdefault.jpg)
![Lite Transformer and Hardware-Aware Transformer, [Microsoft Research, Invited Talk]](https://i.ytimg.com/vi/C-TlROqPVXw/mqdefault.jpg)