Simran Arora (@simran_s_arora) 's Twitter Profile
Simran Arora

@simran_s_arora

cs @StanfordAILab @hazyresearch

ID: 4712264894

linkhttps://arorasimran.com/ calendar_today05-01-2016 06:18:44

311 Tweet

3,3K Takipçi

193 Takip Edilen

Simran Arora (@simran_s_arora) 's Twitter Profile Photo
Simran Arora
@simran_s_arora

12 days ago

There’s been tons of work on KV-cache compression and KV-cache free Transformer-alternatives (SSMs, linear attention) models for long-context, but we know there’s no free lunch with these methods. The quality-memory tradeoffs are annoying. *Is all lost?* Introducing CARTRIDGES:

There’s been tons of work on KV-cache compression and KV-cache free Transformer-alternatives (SSMs, linear attention) models for long-context, but we know there’s no free lunch with these methods. The quality-memory tradeoffs are annoying. *Is all lost?* Introducing CARTRIDGES: