TriAttention: Efficient Long Reasoning with Trigonometric KV Compression
Details
In this session, Damian will be presenting: TriAttention: Efficient Long Reasoning with Trigonometric KV Compression by Weian Mao, Xi Lin, Wei Huang, Yuxin Xie, Tianfu Fu, Bohan Zhuang, Song Han, Yukang Chen
--------------------------------------------------------------------------------
We discuss a different research paper every week. We post each week's paper in our GitHub repo - please read it before the meetup.
All events will be hosted on a Google Meet video call. Occasionally we also host an in-person event in our London office - watch this space for updates.
All recorded presentations can be found in our YouTube channel (don't forget to subscribe!).
Join our Discord channel: https://discord.gg/a2fMv7YAfs
