/CARDS

PyTorch implementation of the paper "Cascade Reward Sampling for Efficient Decoding-Time Alignment"

Primary LanguagePython

This repository is not active