Nexus Prime is a comprehensive multimodal AI framework that fuses text, vision, audio, and real-time data to support zero-shot learning, ethical decision-making, and scenario simulations such as climate modeling and personalized medicine. The server exposes an API (via FastAPI) for real-time inference, along with tools for model simulation, training, and edge deployment. You can run the server with Python by starting the API module, then interact with the /infer endpoint to submit multimodal inputs and receive predictions. The project also includes utilities for ONNX export, distributed training, and VR/AR-enabled simulations, making it suitable for research and production deployments that require scalable, ethically audited AI capabilities.

Prerequisites:

• Python 3.8+
• CUDA-enabled GPU recommended for performance (optional for CPU)
• Docker (optional for containerized deployment)
• Git
• (Optional) Qiskit account for quantum features

Step-by-step installation:

1. Clone the repository: git clone https://github.com/KOSASIH/nexus-prime.git cd nexus-prime

2. Create and activate a virtual environment (recommended): python -m venv venv source venv/bin/activate # Linux/macOS venv\Scripts\activate # Windows

3. Install dependencies: pip install -r requirements.txt

   Key packages include: torch, transformers, qiskit, ray, fastapi, onnx, pytorch-lightning, pyvista, kubernetes, etc.

4. Optional: Set up quantum prerequisites

   • Install Qiskit: pip install qiskit
   • If you have a Qiskit account, load credentials as needed in your config

5. Run the API server (example): python -m nexus_prime.inference.api

   The server will typically start on http://localhost:8000

6. Optional Docker-based deployment: docker build -t nexusprime:latest . docker run -it -p 8000:8000 nexusprime:latest

7. Optional: Download pre-trained weights (if applicable): python src/nexus_prime/utils/download_weights.py --api-key YOUR_KEY

Notes:

• Ensure CUDA drivers and compatible PyTorch version are installed for GPU acceleration.
• If you plan to run distributed training or edge deployments, follow the project’s distributed/edge deployment guidelines in the docs.

Tips and common issues:

• If the API does not start, check that port 8000 is free or configure the server to use a different port.
• For quantum features, you may need a Qiskit account; otherwise a classical fallback is used.
• When using Docker, ensure Docker daemon is running and you have enough memory for large models.
• Set environment variables like NEXUS_API_KEY only if you need authenticated access to weight files or external services.
• For inference performance, consider exporting to ONNX and deploying on edge hardware with the provided edge tools.
• If you encounter dependency conflicts, consider using a clean virtual environment and installing pinned versions from requirements.txt.