AjatusCore

AjatusCore is the heart of the Ajatuskumppani project. It contains the core language model (LLM), fine-tuning scripts, and the inference engine.

Features

• Base Model: Built upon Mistral 7B Instruct
• Fine-tuning: Includes scripts for fine-tuning the model with custom data.
• Inference: Optimized for performance with vLLM and Ollama.
• GGUF Support: Ready for on-device AI with GGUF.

Getting Started

Prerequisites

• Python 3.11+
• PyTorch 2.0+
• CUDA 11.8+

Installation

1. Clone the repository:

   git clone https://github.com/pinnacore-ai/ajatuskumppani.git
   cd ajatuskumppani/ajatus-core

2. Install the dependencies:

   pip install -r requirements.txt

Usage

Inference

To run inference with the model, use the following command:

python inference/run.py --prompt "Hello, world!"

Fine-tuning

To fine-tune the model with your own data, see the instructions in the training directory.

License

This project is licensed under the Apache 2.0 License. See the LICENSE file for details.