pn-summary-mt5-base Open-source Model - Accurately Generate High-quality Persian Text Summaries

Pn Summary Mt5 Base

Developed by HooshvareLab

This model is a Persian text summarization model based on the mT5-base architecture, specifically optimized for the pn_summary dataset, capable of generating high-quality Persian text summaries.

Text Generation

Transformers

Other#Persian summarization #Multilingual MT5 #High ROUGE score

Downloads 68

Release Time : 3/2/2022

Model Overview

This model is primarily used for Persian text summarization tasks, extracting key information from long texts to generate concise summaries.

Model Features

Persian optimization

Specifically optimized for Persian text, better handling the linguistic characteristics of Persian

Multi-granularity summarization

Supports generating summaries of different lengths and detail levels (low, medium, high)

High performance

Excellent performance on both validation and test sets, with ROUGE-1 F1 score approaching 50%

Model Capabilities

Persian text comprehension

Key information extraction

Multi-length summary generation

Long text processing

Use Cases

News media

News summarization

Automatically generate concise summaries of Persian news articles

Saves editorial time and improves content distribution efficiency

Academic research

Paper summarization

Generate structured summaries for Persian academic papers

Helps researchers quickly grasp the core content of papers

🚀 mT5-base Model for Article Summarization

This is an mT5-base model trained on the pnSummary dataset, designed to summarize articles effectively.

✨ Features

Language and Tags: This model is related to Persian language (fa), with tags including summarization and mt5. Its pipeline tag is summarization.
Dataset: It is trained on the pnSummary dataset.
Widget Example: Here is an example text for summarization: "Bob Odenkirk, a 58 - year - old American actor who started his career as a comedy writer in the late 1980s and gained wider fame through his appearance in the TV series 'Breaking Bad', has recently attracted more media attention than ever in the lead role of an action movie. Odenkirk began his activities as a writer for 'Saturday Night Live' in 1987. His work as a writer was followed by playing several small roles. Odenkirk showed his acting talent from the start and in the 1990s, he appeared as a writer and actor in 'The Ben Stiller Show' in pursuit of his dream. The peak of Odenkirk's acting in TV sitcoms was marked by 'Mr. Show with Bob and David', although his career continued with small roles in shows like 'Seinfeld' and 'How I Met Your Mother'. In 2009, when Bob Odenkirk joined the cast of 'Breaking Bad' to play the role of a corrupt lawyer named Saul Goodman, he experienced a major turning point in his professional life. Playing Saul in 'Breaking Bad' paved the way for him to play the lead role in 'Better Call Saul', which was probably the most important role of Odenkirk's life."

📚 Documentation

Eval results

The following table summarizes the ROUGE scores obtained by the model for the validation set.

+-----------+------+-----------+--------+-----------+
|   Score   | Type | Precision | Recall | F-Measure |
+-----------+------+-----------+--------+-----------+
|   rouge1  | low  |   48.60   | 45.25  |   45.47   |
|   rouge1  | mid  |   49.22   | 45.79  |   46.02   |
|   rouge1  | high |   49.80   | 46.36  |   46.54   |
|           |      |           |        |           |
|    ***    | ***  |    ***    |  ***   |    ***    |
|           |      |           |        |           |
|   rouge2  | low  |   29.46   | 27.33  |   27.56   |
|   rouge2  | mid  |   30.08   | 27.93  |   28.12   |
|   rouge2  | high |   30.69   | 28.52  |   28.71   |
|           |      |           |        |           |
|    ***    | ***  |    ***    |  ***   |    ***    |
|           |      |           |        |           |
|   rougeL  | low  |   42.20   | 39.26  |   39.51   |
|   rougeL  | mid  |   42.80   | 39.86  |   40.07   |
|   rougeL  | high |   43.43   | 40.41  |   40.62   |
|           |      |           |        |           |
|    ***    | ***  |    ***    |  ***   |    ***    |
|           |      |           |        |           |
| rougeLsum | low  |   42.15   | 39.26  |   39.48   |
| rougeLsum | mid  |   42.82   | 39.87  |   40.09   |
| rougeLsum | high |   43.39   | 40.43  |   40.61   |
+-----------+------+-----------+--------+-----------+

Test results

The following table summarizes the ROUGE scores obtained by the model for the test set.

+-----------+------+-----------+--------+-----------+
|   Score   | Type | Precision | Recall | F-Measure |
+-----------+------+-----------+--------+-----------+
|   rouge1  | low  |   48.83   | 44.99  |   45.41   |
|   rouge1  | mid  |   49.39   | 45.52  |   45.91   |
|   rouge1  | high |   50.02   | 46.07  |   46.48   |
|           |      |           |        |           |
|    ***    | ***  |    ***    |  ***   |    ***    |
|           |      |           |        |           |
|   rouge2  | low  |   29.27   | 26.90  |   27.23   |
|   rouge2  | mid  |   29.97   | 27.52  |   27.84   |
|   rouge2  | high |   30.62   | 28.15  |   28.46   |
|           |      |           |        |           |
|    ***    | ***  |    ***    |  ***   |    ***    |
|           |      |           |        |           |
|   rougeL  | low  |   42.30   | 38.94  |   39.34   |
|   rougeL  | mid  |   42.85   | 39.54  |   39.90   |
|   rougeL  | high |   43.51   | 40.12  |   40.49   |
|           |      |           |        |           |
|    ***    | ***  |    ***    |  ***   |    ***    |
|           |      |           |        |           |
| rougeLsum | low  |   42.27   | 38.94  |   39.35   |
| rougeLsum | mid  |   42.86   | 39.53  |   39.90   |
| rougeLsum | high |   43.42   | 40.12  |   40.45   |
+-----------+------+-----------+--------+-----------+

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