This repository contains datasets for training and evaluating machine learning models for a chat agent. The datasets cover various domains, including room math, MMLU, reasoning, and additional benchmarks. The goal is to achieve 100% accuracy, high quality, quantity, and lack of bias in the datasets.
To train models using these datasets and evaluate their performance against the benchmarks, follow these steps:
- Install the required dependencies:
pip install -r requirements.txt
- Train the model using the
train_model.py
script:python train_model.py --dataset combined_dataset.csv
- Evaluate the model using the
evaluate_model.py
script:python evaluate_model.py --dataset combined_dataset.csv --model_path path_to_trained_model
Due to the size limitations of Git, the large dataset files are not included directly in the repository. Instead, they are available as assets in the GitHub release. You can download the large files from the following links:
Contributions are welcome! Please follow the GitHub Pull Request Workflow to contribute to this repository. To contribute:
- Fork the repository.
- Create a new branch for your changes.
- Make your changes and commit them with clear and descriptive messages.
- Push your changes to your forked repository.
- Create a pull request to the main repository.
This repository is licensed under the MIT License. See the LICENSE file for more information.