HazyResearch/based

Code for exploring Based models from "Simple linear attention language models balance the recall-throughput tradeoff"

PythonApache-2.0