Implementation of Griffin from the paper: "Griffin: Mixing Gated Linear Recurrences with Local Attention for Efficient Language Models"
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).