TLDR · QUICK LINKS
IS ATTENTION ALL YOU NEED?
Summary from the TLDR newsletter
Several new architectures have recently emerged that are designed to outperform full attention Transformers in long-context learning/generation and inference speed/cost.
TLDR · QUICK LINKS
Several new architectures have recently emerged that are designed to outperform full attention Transformers in long-context learning/generation and inference speed/cost.