BlenderBot Small-90M Open-Source Conversation Model - Supports Multi-turn Natural Conversations and Integrates Diverse Conversation Skills

Blenderbot Small 90M

Developed by facebook

This is an open-domain dialogue system based on large-scale neural networks, capable of conducting multi-turn natural conversations and integrating various dialogue skills.

Dialogue System EnglishOpen Source License:Apache-2.0 #Multi-turn dialogue generation #Large-scale parameter model #Hybrid skill fusion

Downloads 2,407

Release Time : 3/2/2022

Model Overview

This model focuses on open-domain dialogue generation, providing multiple dialogue skills such as topic provision, Q&A, knowledge display, and empathy expression. Developed based on the paper 'Recipes for Building an Open-Domain Chatbot,' it offers model variants with different parameter scales.

Model Features

Multi-skill fusion

Capable of integrating various dialogue skills such as questioning, answering, knowledge display, and empathy expression.

Large-scale model

Offers model choices ranging from 90 million to 9.4 billion parameters.

Human-like dialogue

Outperforms existing methods in attractiveness and human-likeness metrics for multi-turn dialogues.

Model Capabilities

Open-domain dialogue generation

Multi-turn dialogue maintenance

Topic guidance

Empathy expression

Knowledge display

Use Cases

Intelligent customer service

Customer service dialogue

Used for handling customer inquiries and providing assistance in natural dialogue scenarios.

Delivers a more natural and fluent customer service experience.

Social companionship

Virtual companionship dialogue

Acts as a virtual companion for daily conversation exchanges.

Provides a more engaging and human-like dialogue experience.

🚀 Chatbot Model for Open - domain Conversation

This project provides open - domain chatbot models. By using appropriate training data and generation strategies, large - scale models can learn various conversation skills, and the models outperform existing approaches in multi - turn dialogue.

🚀 Quick Start

No specific quick - start steps are provided in the original document.

✨ Features

Skill Learning: Large - scale models can learn a variety of conversation skills such as providing engaging talking points, listening, asking and answering questions, and displaying knowledge, empathy and personality when given appropriate training data and generation strategies.
Public Availability: The models and code with 90M, 2.7B and 9.4B parameters are publicly available.
Superior Performance: Human evaluations show that the best models are superior to existing approaches in multi - turn dialogue in terms of engagingness and humanness measurements.

📚 Documentation

Model description

Paper: Recipes for building an open - domain chatbot
Original PARLAI Code: Original PARLAI Code

Abstract

Building open - domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we show that other ingredients are important for a high - performing chatbot. Good conversation requires a number of skills that an expert conversationalist blends in a seamless way: providing engaging talking points and listening to their partners, both asking and answering questions, and displaying knowledge, empathy and personality appropriately, depending on the situation. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter neural models, and make our models and code publicly available. Human evaluations show our best models are superior to existing approaches in multi - turn dialogue in terms of engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.

📄 License

The model is licensed under the apache - 2.0 license.

Property	Details
Model Type	Not specified
Training Data	blended_skill_talk
Metrics	perplexity

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご