MatchSum

Visit Tool

MatchSum is a Research & Education tool that performs extractive summarization as text matching. It provides code for the ACL 2020 paper and supports various datasets for summarization tasks.

Claim this tool

No Views Yet

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is MatchSum?

MatchSum offers an implementation of the ACL 2020 paper "Extractive Summarization as Text Matching." This tool is designed for researchers and developers working on natural language processing tasks, specifically extractive summarization. It supports both BERT and RoBERTa encoders and provides pre-trained models for the CNN/DailyMail dataset, as well as other datasets like WikiHow, PubMed, XSum, MultiNews, and Reddit. Users can process their own data by converting it to a specific JSONL format and generating candidate summaries. The code requires Python 3.7, PyTorch 1.4.0, fastNLP 0.5.0, pyrouge 0.1.3, rouge 1.0.0, and transformers 2.5.1, and is optimized for Linux environments with GPU support.

Best used for

Ideal for researchers and students who need to implement and evaluate state-of-the-art extractive summarization models, process large text datasets, and experiment with different encoder architectures. Especially valuable for academic projects and NLP research focusing on text matching techniques.

Common actions

summarize text

extract key information

process natural language

collaborationlow-code/no-codedeepfakeworkflowsautomated workflowface swapping"AI Agents"github copilotopen-source

Capabilities

Key features

Extractive summarization
Text matching approach
BERT/RoBERTa encoders
Pre-trained models
Custom data processing
ROUGE score evaluation

Target Audience

professorstudent

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

What are the system requirements for running MatchSum?

MatchSum requires Python 3.7, PyTorch 1.4.0, fastNLP 0.5.0, pyrouge 0.1.3, rouge 1.0.0, and transformers 2.5.1. It is designed to run on Linux operating systems and benefits from GPU acceleration, specifically tested with Tesla-V100-16G GPUs.

Can I use MatchSum with my own datasets?

Yes, you can process your own data. You need to convert your dataset to a JSONL format with 'text' and 'summary' fields. Additionally, you'll need to use a method like BertExt to select important sentences and create an index.jsonl file before generating candidates.

Are there pre-trained models available for MatchSum?

Yes, MatchSum provides pre-trained models for the CNN/DailyMail dataset, available in both BERT and RoBERTa versions. Pre-trained models for other datasets like WikiHow, PubMed, XSum, MultiNews, and Reddit are also provided.

Trending

Subcategories trending in Research & Education

Academic Research Study Assistants Knowledge Management Course Creation Scientific Computing Language Learning

Trending

Also listed in

This tool also appears in

AI Agents & Automation › RAG & Document AI

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce