Skip to content

Latest commit

 

History

History
79 lines (54 loc) · 2.58 KB

File metadata and controls

79 lines (54 loc) · 2.58 KB

🎖️ WarMatrix: Universal Setup Guide

Welcome to the WarMatrix Command Center. Follow this guide to initialize the tactical simulation suite and prepare for deployment.


📋 Phase 1: System Prerequisites

Before proceeding, ensure your station is equipped with the following:

  • Node.js: v20 or higher.
  • Python: v3.10 to v3.12 (v3.12 recommended).
  • Git: Required for repository management and model acquisition.
  • NVIDIA GPU: Required for high-performance local AI inference (VRAM: 8GB+ recommended, 4GB minimum with quantization).

🔌 Phase 2: Core Application Setup

Initialize the primary interface and the simulation engine.

  1. Frontend Installation:

    # Install dependencies
    npm install
  2. Simulation Backend Setup:

    # Navigate to root (if not already there)
    cd WarMatrix
    
    # Create and activate virtual environment
    python -m venv .venv
    & .\.venv\Scripts\Activate.ps1
    
    # (1) Install core simulation engine requirements
    #     Installs: numpy, fastapi, uvicorn, pydantic
    pip install -r backend/requirements.txt
    
    # (2) Install AI server requirements (GPU/LLM inference stack)
    #     Installs: torch (CUDA 12.8), transformers, peft, bitsandbytes, accelerate
    #     NOTE: Skip this step if you are using LM Studio (Method C in Phase 3).
    pip install -r ai_server/requirements.txt

🧠 Phase 3: AI Server Configuration

WarMatrix uses a dedicated AI server for tactical synthesis and scenario generation. Choose the method that best fits your hardware capabilities:

Method A: Quick Start / Standard (Recommended)

Use this if you want to get up and running quickly using a standard virtual environment. 👉 View Quick Start Guide

Method B: High-Performance GPU (Conda/Unsloth)

Use this for the best performance. Optimized for NVIDIA GPUs using Conda and Unsloth for ultra-fast response times. 👉 View High-Performance Guide

Method C: LM Studio Proxy (Lightweight)

Use this if you are VRAM-constrained or want to run the model on a separate machine in your network via LM Studio. 👉 View LM Studio Guide


🚀 Phase 4: Launching the System

Once all components are configured, you can launch the entire suite concurrently:

# Launch both Frontend and Backend concurrently
npm run dev

The Command Console will be available at http://localhost:3000.


Document maintained by the WarMatrix Operational Command.