twitter-xlm-roberta-base-sentiment-finetunned Open Source Model - Free Implementation of Multilingual Text Sentiment Analysis

Twitter Xlm Roberta Base Sentiment Finetunned

Developed by citizenlab

A multilingual XLM-Roberta sequence classification model for text sentiment analysis, fine-tuned based on the sentiment classification model from the Cardiff NLP team.

Text Classification

Transformers

Supports Multiple Languages#Multilingual Sentiment Analysis #Social Media Text Classification #Toxic Content Detection

Downloads 10.41k

Release Time : 3/2/2022

Model Overview

This model is a multilingual sentiment classification model capable of identifying positive, neutral, and negative sentiments in text. Based on the XLM-Roberta architecture and fine-tuned on multiple languages.

Model Features

Multilingual Support

Supports sentiment analysis in 10 languages including English, French, German, etc.

High Accuracy

Achieves 0.80 accuracy on the test set, with an F1 score of 0.85 for the positive class

Fine-tuned Pre-trained Model

Fine-tuned based on the powerful XLM-Roberta pre-trained model, with excellent text understanding capabilities

Model Capabilities

Text Sentiment Classification

Multilingual Text Analysis

Toxic Content Detection

Use Cases

Social Media Analysis

Tweet Sentiment Analysis

Analyze user sentiment tendencies on social media platforms like Twitter

Can accurately identify positive, neutral, and negative sentiments

Content Moderation

Harmful Content Detection

Identify negative content containing insults, hate speech, etc.

57% accuracy in detecting negative content

🚀 citizenlab/twitter-xlm-roberta-base-sentiment-finetunned

This is a multilingual XLM - Roberta model sequence classifier. It's fine - tuned and based on the Cardiff NLP Group sentiment classification model, aiming to provide accurate sentiment classification for multiple languages.

🚀 Quick Start

✨ Features

Multilingual Support: Supports multiple languages including English (en), Dutch (nl), French (fr), Portuguese (pt), Italian (it), Spanish (es), German (de), Danish (da), Polish (pl), and Afrikaans (af).
Text Classification Pipeline: It can be used for text - classification tasks, especially sentiment analysis.

📦 Installation

There is no specific installation command provided in the original document. However, to use the model, you need to have the transformers library installed. You can install it using the following command:

pip install transformers

💻 Usage Examples

Basic Usage

from transformers import pipeline

model_path = "citizenlab/twitter-xlm-roberta-base-sentiment-finetunned"

sentiment_classifier = pipeline("text-classification", model=model_path, tokenizer=model_path)
sentiment_classifier("this is a lovely message")
> [{'label': 'Positive', 'score': 0.9918450713157654}]

sentiment_classifier("you are an idiot and you and your family should go back to your country")
> [{'label': 'Negative', 'score': 0.9849833846092224}]

📚 Documentation

Model Information

Property	Details
Pipeline Type	text - classification
Supported Languages	en, nl, fr, pt, it, es, de, da, pl, af
Datasets	jigsaw_toxicity_pred
Metrics	F1 Accuracy

Example Widgets

The model comes with some example widgets for quick testing:

Example 1:
- Text: "this is a lovely message"
- Multi - Class: false
Example 2:
- Text: "you are an idiot and you and your family should go back to your country"
- Multi - Class: false

🔧 Technical Details

Evaluation

              precision    recall  f1-score   support

    Negative       0.57      0.14      0.23        28
     Neutral       0.78      0.94      0.86       132
    Positive       0.89      0.80      0.85        51

    accuracy                           0.80       211
   macro avg       0.75      0.63      0.64       211
weighted avg       0.78      0.80      0.77       211

This evaluation shows the performance of the model on different sentiment classes (Negative, Neutral, Positive) in terms of precision, recall, f1 - score, and support.

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