albert-base-v2-emotion Open-source Sentiment Analysis Model - Lightweight and Accurate Analysis of Twitter Sentiment

Albert Base V2 Emotion

Developed by bhadresh-savani

Lightweight sentiment analysis model based on ALBERT architecture, fine-tuned on Twitter sentiment dataset

Text Classification EnglishOpen Source License:Apache-2.0 #Lightweight BERT #Sentiment Classification #Twitter Analysis

Downloads 15.44k

Release Time : 3/2/2022

Model Overview

This model is a lightweight text classification model based on the ALBERT-base-v2 architecture, specifically designed for sentiment analysis tasks. It has been fine-tuned on a Twitter sentiment dataset and can identify six basic emotions in text: sadness, joy, love, anger, fear, and surprise.

Model Features

Lightweight Architecture

Uses ALBERT architecture with significantly fewer parameters than traditional BERT models while maintaining high performance

Efficient Sentiment Analysis

Optimized specifically for sentiment analysis tasks, accurately identifying six basic emotions

Fast Inference

Capable of processing approximately 183 test samples per second, suitable for real-time applications

Model Capabilities

Text Classification

Sentiment Analysis

Emotion Recognition

Use Cases

Social Media Analysis

Tweet Sentiment Analysis

Analyzing user sentiment trends on social media platforms like Twitter

Achieved 93.6% accuracy on the test set

Customer Feedback Analysis

Product Review Sentiment Classification

Automatically classifying sentiment tendencies in customer reviews

Effectively identifies key emotions such as joy and anger

🚀 Albert-base-v2-emotion

This is a text classification model based on the Albert-base-v2 architecture, fine - tuned on the emotion dataset for emotion classification tasks.

🚀 Quick Start

Usage Example

from transformers import pipeline
classifier = pipeline("text-classification",model='bhadresh-savani/albert-base-v2-emotion', return_all_scores=True)
prediction = classifier("I love using transformers. The best part is wide range of support and its easy to use", )
print(prediction)

"""
Output:
[[
{'label': 'sadness', 'score': 0.010403595864772797}, 
{'label': 'joy', 'score': 0.8902180790901184}, 
{'label': 'love', 'score': 0.042532723397016525}, 
{'label': 'anger', 'score': 0.041297927498817444}, 
{'label': 'fear', 'score': 0.011772023513913155}, 
{'label': 'surprise', 'score': 0.0037756056990474463}
]]
"""

✨ Features

Lite Architecture: Albert is a Lite BERT architecture with significantly fewer parameters than a traditional BERT architecture.
Fine - tuned on Emotion Dataset: Albert - base - v2 is fine - tuned on the emotion dataset using HuggingFace Trainer.

📚 Documentation

Model description

Albert is A Lite BERT architecture that has significantly fewer parameters than a traditional BERT architecture.

Albert - base - v2 is fine - tuned on the emotion dataset using HuggingFace Trainer with the following hyperparameters:

learning rate 2e - 5, 
batch size 64,
num_train_epochs = 8,

Model Performance Comparision on Emotion Dataset from Twitter

Model	Accuracy	F1 Score	Test Sample per Second
Distilbert - base - uncased - emotion	93.8	93.79	398.69
Bert - base - uncased - emotion	94.05	94.06	190.152
Roberta - base - emotion	93.95	93.97	195.639
Albert - base - v2 - emotion	93.6	93.65	182.794

Dataset

Twitter - Sentiment - Analysis.

Training procedure

Colab Notebook

Eval results

{
'test_accuracy': 0.936,
 'test_f1': 0.9365658988006296,
 'test_loss': 0.15278364717960358,
 'test_runtime': 10.9413,
 'test_samples_per_second': 182.794,
 'test_steps_per_second': 2.925
 }

Reference

Natural Language Processing with Transformer By Lewis Tunstall, Leandro von Werra, Thomas Wolf

📄 License

This model is licensed under the Apache - 2.0 license.

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