This project has been archived. If you need to use this project, please revert to ParlAI 1.4.1.

Multi-Modal Open-Domain Dialogue

Kurt Shuster, Eric Michael Smith, Da Ju, Jason Weston


Recent work in open-domain conversational agents has demonstrated that significant improvements in model engagingness and humanness metrics can be achieved via massive scaling in both pre-training data and model size (Adiwardana et al., 2020; Roller et al., 2020). However, if we want to build agents with human-like abilities, we must expand beyond handling just text. A particularly important topic is the ability to see images and communicate about what is perceived. With the goal of engaging humans in multi-modal dialogue, we investigate combining components from state-of-the-art open-domain dialogue agents with those from state-of-the-art vision models. We study incorporating different image fusion schemes and domain-adaptive pre-training and fine-tuning strategies, and show that our best resulting model outperforms strong existing models in multi-modal dialogue while simultaneously performing as well as its predecessor (text-only) BlenderBot (Roller et al., 2020) in text-based conversation.



Example conversations

Code and models

See the model card for details.


If you use the models in your own work, please cite with the following BibTex entry:

      title={Multi-Modal Open-Domain Dialogue}, 
      author={Kurt Shuster and Eric Michael Smith and Da Ju and Jason Weston},