Mixture-of-Depths: Dynamically allocating compute in transformer-based language models

Add code
Apr 02, 2024

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: