Implementation of Memorizing Transformers (ICLR 2022), attention net augmented with indexing and retrieval of memories using approximate nearest neighbors, in Pytorch
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).