bert-base-cased-ag-news Open-Source Text Classification Model - Free Deployment for 94.5% Accuracy in News Classification

Home

Bert Base Cased Ag News

Developed by lucasresck

A BERT-based text classification model, fine-tuned on the AG News dataset with a test accuracy of 94.5%.

Text Classification

Transformers

EnglishOpen Source License:MIT #News Classification #High Accuracy BERT #English Text Processing

Downloads 150

Release Time : 3/2/2022

Model Overview

This model adds a linear layer on top of BERT's [CLS] token output, specifically designed for news text classification tasks.

Model Features

High Accuracy

Achieves 94.5% accuracy on the AG News test set.

BERT Fine-Tuning

Fine-tuned based on the pre-trained BERT model, fully leveraging pre-trained language representations.

Multi-Class Classification

Supports classification into 4 news categories (World, Sports, Business, Technology).

Model Capabilities

English Text Classification

News Category Prediction

Short Text Understanding

Use Cases

News Classification

Automatic News Classification

Automatically classify news articles into predefined categories.

94.5% accuracy

Content Filtering

Filter specific types of news content based on categories.

🚀 bert-base-cased-ag-news

A BERT model fine-tuned on the AG News classification dataset. It uses a linear layer on top of the [CLS] token output and achieves a test accuracy of 0.945.

🚀 Quick Start

Basic Usage

# Here is how to use this model to classify a given text:
from transformers import AutoTokenizer, BertForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained('lucasresck/bert-base-cased-ag-news')
model = BertForSequenceClassification.from_pretrained('lucasresck/bert-base-cased-ag-news')
text = "Is it soccer or football?"
encoded_input = tokenizer(text, return_tensors='pt', truncation=True, max_length=512)
output = model(**encoded_input)

✨ Features

Fine-tuned on AG News: The model is fine-tuned on the AG News classification dataset.
High Accuracy: Achieves a test accuracy of 0.945.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

# Here is how to use this model to classify a given text:
from transformers import AutoTokenizer, BertForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained('lucasresck/bert-base-cased-ag-news')
model = BertForSequenceClassification.from_pretrained('lucasresck/bert-base-cased-ag-news')
text = "Is it soccer or football?"
encoded_input = tokenizer(text, return_tensors='pt', truncation=True, max_length=512)
output = model(**encoded_input)

📚 Documentation

Limitations and bias

Bias were not assessed in this model, but, considering that pre-trained BERT is known to carry bias, it is also expected for this model. BERT's authors say: "This bias will also affect all fine-tuned versions of this model."

Evaluation results

              precision    recall  f1-score   support

           0     0.9539    0.9584    0.9562      1900
           1     0.9884    0.9879    0.9882      1900
           2     0.9251    0.9095    0.9172      1900
           3     0.9127    0.9242    0.9184      1900

    accuracy                         0.9450      7600
   macro avg     0.9450    0.9450    0.9450      7600
weighted avg     0.9450    0.9450    0.9450      7600

📄 License

This project is licensed under the MIT license.

Property	Details
Model Type	BERT fine-tuned for AG News classification
Training Data	AG News classification dataset
Metrics	accuracy, f1, recall, precision

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