bart-finetuned-samsum Open-Source Dialogue Summarization Model - Optimize Dialogue Summarization Effect, Free to Use!

Bart Finetuned Samsum

Developed by luisotorres

A dialogue summarization model fine-tuned based on BART-large-xsum, specifically optimized for the SamSum dialogue dataset

Text Generation

Transformers

English#Dialogue Summarization #High ROUGE Score #English Conversation Processing

Downloads 177

Release Time : 10/30/2023

Model Overview

This model specializes in generating summaries of dialogue texts, capable of extracting key information from conversations and producing concise summaries.

Model Features

Dialogue Summarization Optimization

Fine-tuned specifically for dialogue data, effectively understanding dialogue context and generating accurate summaries

Based on BART Architecture

Utilizes BART's powerful sequence-to-sequence learning capability to achieve high-quality text generation

SamSum Dataset Fine-tuning

Trained using the professional dialogue summarization dataset SamSum to enhance performance in dialogue scenarios

Model Capabilities

Dialogue text summarization

Key information extraction

Natural language generation

Use Cases

Dialogue Analysis

Customer Service Dialogue Summary

Automatically generates summaries of key issues and solutions from customer service dialogues

Improves customer service efficiency and facilitates subsequent analysis

Meeting Minutes Condensation

Converts lengthy meeting dialogues into concise key-point summaries

Saves reading time and quickly captures the core content of meetings

🚀 bart-finetuned-samsum

A fine - tuned model for dialogue summarization, based on the facebook/bart - large - xsum architecture and optimized using the SamSum dataset.

🚀 Quick Start

This model is designed for text summarization. You can use it as follows:

from transformers import pipeline

model = pipeline("summarization", model="luisotorres/bart-finetuned-samsum")

conversation = '''Sarah: Do you think it's a good idea to invest in Bitcoin?
    Emily: I'm skeptical. The market is very volatile, and you could lose money.
    Sarah: True. But there's also a high upside, right?                                     
'''
model(conversation)

✨ Features

Specialized adaptation of the facebook/bart-large-xsum for dialogue summarization.
Fine - tuned on the SamSum dataset to enhance performance in summarizing dialogues.

📦 Installation

The README does not provide specific installation steps, so this section is skipped.

💻 Usage Examples

Basic Usage

from transformers import pipeline

model = pipeline("summarization", model="luisotorres/bart-finetuned-samsum")

conversation = '''Sarah: Do you think it's a good idea to invest in Bitcoin?
    Emily: I'm skeptical. The market is very volatile, and you could lose money.
    Sarah: True. But there's also a high upside, right?                                     
'''
model(conversation)

Advanced Usage

The README does not provide advanced usage examples, so this part is skipped.

📚 Documentation

Model Information

Property	Details
Model Type	bart-finetuned-samsum
Training Data	SamSum
Pipeline Tag	summarization
Metrics	rouge
Library Name	transformers

Model Performance

The model bart-finetuned-samsum has the following performance metrics on the SamSum dataset:

Task	Dataset	Metric	Value
Text Summarization	SamSum	Validation ROUGE - 1	53.8804
Text Summarization	SamSum	Validation ROUGE - 2	29.2329
Text Summarization	SamSum	Validation ROUGE - L	44.774
Text Summarization	SamSum	Validation ROUGE - L Sum	49.8255
Text Summarization	SamSum	Test ROUGE - 1	52.8156
Text Summarization	SamSum	Test ROUGE - 2	28.1259
Text Summarization	SamSum	Test ROUGE - L	43.7147
Text Summarization	SamSum	Test ROUGE - L Sum	48.5712

Development

Kaggle Notebook: Text Summarization with Large Language Models

Training Parameters

evaluation_strategy = "epoch",
save_strategy = 'epoch',
load_best_model_at_end = True,
metric_for_best_model = 'eval_loss',
seed = 42,
learning_rate=2e-5,
per_device_train_batch_size=4,
per_device_eval_batch_size=4,
gradient_accumulation_steps=2,
weight_decay=0.01,
save_total_limit=2,
num_train_epochs=4,
predict_with_generate=True,
fp16=True,
report_to="none"

Reference

This model is based on the original BART architecture, as detailed in:

Lewis et al. (2019). BART: Denoising Sequence - to - Sequence Pre - training for Natural Language Generation, Translation, and Comprehension. arXiv:1910.13461

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