/dilated-self-attention

Implementation of the dilated self attention as described in "LongNet: Scaling Transformers to 1,000,000,000 Tokens"

Primary LanguagePythonMIT LicenseMIT

Stargazers

No one’s star this repository yet.