microsoft/DialogRPT-human-vs-machine

DialogRPT is a set of dialog response ranking models proposed by Microsoft Research NLP Group trained on 100 + millions of human feedback data . The human_vs_machine score predicts how likely the response is from a human rather than a machine . It can be used to improve existing dialog generation model (e.g., DialoGPT) by re-ranking the generated response candidates . We considered the following tasks and provided corresponding pretrained models for each task . The Colab Notebook Demo is available to download and use in the next week's version of this article . We hope to use it in the future to improve DialoRPT's ability to rank dialog responses in a new,