tabulargrpo 0.1.3

A TabularGRPO classifier package

TabularGRPO

Overview

TabularGRPO

Dataset

Dataset download url: https://zenodo.org/records/15081413. After you download, please drop in data/ folder

Project Structure

The project structure is as follows:

• tabulargrpo: Root directory
  • README.md: the reamde
  • data: Data directory
    • synthetic_data_small.csv: Download dataset and drop here
  • models: Model directory
    • MoETransformer.py: MoETransformer model
  • tabulargrpo_classifier.py: Training script
  • demo.py: Demo script

Features

• Tabular Classifier: Tabular Data Training
• MoE Transformer: MoE Transformer
• GRPO: GRPO

Installation

To install the TabularGRPO project, follow these steps:

1. Clone the repository:

   git clone https://github.com/yourusername/tabulargrpo.git
   

2. Navigate to the project directory:

   cd tabulargrpo
   

3. Install the required dependencies:

   pip install -r requirements.txt
   

Usage

To use the TabularGRPO train, run the following command:

        from tabulargrpo_classifier import TabularGRPOClassifier
        from models.transformer_moe import MoETransformer
        from sklearn.datasets import make_classification
        from sklearn.model_selection import train_test_split
        
        X, y = make_classification(n_samples=5000, n_features=14, n_classes=2, random_state=42)
 
        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
        
        classifier = TabularGRPOClassifier(model_class=MoETransformer,input_dim=14, num_classes=2, epochs=10,group_size=10)
        classifier.fit(X_train,y_train)

To use the TabularGRPO evaluate, run the following command:

        from tabulargrpo_classifier import TabularGRPOClassifier
        from models.transformer_moe import MoETransformer
        from sklearn.datasets import make_classification
        from sklearn.model_selection import train_test_split
        
        X, y = make_classification(n_samples=5000, n_features=14, n_classes=2, random_state=42)
 
        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
        
        classifier = TabularGRPOClassifier(model_class=MoETransformer,input_dim=14, num_classes=2, epochs=10,group_size=10)
        classifier.fit(X_train,y_train)
        classifier.evaluate(X_test, y_test)

To use the TabularGRPO predict, run the following command:

        from tabulargrpo_classifier import TabularGRPOClassifier
        from models.transformer_moe import MoETransformer
        from sklearn.datasets import make_classification
        from sklearn.model_selection import train_test_split
        
        X, y = make_classification(n_samples=5000, n_features=14, n_classes=2, random_state=42)
 
        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
        
        classifier = TabularGRPOClassifier(model_class=MoETransformer,input_dim=14, num_classes=2, epochs=10,group_size=10)
        classifier.fit(X_train,y_train)
        classifier.evaluate(X_test, y_test)
        data = [[0,0,0,0]]
        p = classifier.predict(data)
        print(p)

To use the TabularGRPO demo, run the following command:

        python demo.py

Contributing

We welcome contributions to the TabularGRPO project. To contribute, please follow these steps:

1. Fork the repository.
2. Create a new branch for your feature or bugfix.
3. Commit your changes and push to your branch.
4. Create a pull request with a detailed description of your changes.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Contact

For any questions or inquiries, please contact us at [e-mail].