RedNinja6440/verilog-slm-krackhack26

⚡ Sovereign AI Verilog Generator

A locally-running AI-powered Verilog code generator with a Cyberpunk-themed interface. Built for the KrackHack Hackathon.

100% Sovereign — No cloud APIs. No data leaves your machine. Runs entirely on your local hardware.

[image]

🚀 Features

• —🧠 Local LLM Inference — Uses the smolify/smolified-krackhack26verilog model (270M params, Gemma3-based)
• —⚡ Cyberpunk UI — Dark theme with neon accents, terminal-style code output, JetBrains Mono font
• —📋 One-Click Copy — Copy generated Verilog to clipboard instantly
• —🔒 Privacy First — All processing happens locally on your machine
• —🎨 Auto-Formatter — Outputs properly indented, multi-line Verilog code

📦 Tech Stack

🛠️ Setup & Installation

Prerequisites

• —Python 3.8+
• —pip

1. Clone the Repository

git clone https://github.com/YOUR_USERNAME/sovereign-verilog-generator.git
cd sovereign-verilog-generator

2. Install Dependencies

pip install -r requirements.txt

3. Start the Server

python3 server.py

⚡ The model will download automatically on first run (~540MB). The server starts at http://localhost:5000.

4. Open the Frontend

Open index.html in your browser — or simply:

xdg-open index.html   # Linux
open index.html        # macOS

💡 Usage

1. 1.Type a circuit description in the Prompt Buffer (left panel)
2. 2.Example: "I need a full adder with inputs A, B, Carry_In and outputs Sum, Carry_Out"
3. 3.Click INITIATE GENERATION_SEQUENCE
4. 4.Wait ~15-30s for the model to generate code
5. 5.The formatted Verilog appears in the Output Stream (right panel)
6. 6.Click COPY to copy the code to your clipboard

📁 Project Structure

├── index.html          # Frontend (single-file, Cyberpunk UI)
├── server.py           # Flask backend with model inference
├── requirements.txt    # Python dependencies
├── start.sh            # Quick-start script
└── README.md           # This file

⚙️ Configuration

Edit server.py to adjust:

• —DEVICE — Change to "cuda" if your GPU supports it (may require debugging)
• —max_new_tokens — Controls output length (default: 512)
• —temperature — Controls randomness (default: 0.7)

🏗️ Architecture

┌─────────────┐     POST /generate      ┌──────────────┐
│  index.html │ ──────────────────────▶  │  server.py   │
│  (Browser)  │                          │  (Flask)     │
│             │ ◀──────────────────────  │              │
│  Tailwind   │     JSON { code: ... }   │  Transformers│
│  + Vanilla  │                          │  + PyTorch   │
│    JS       │                          │              │
└─────────────┘                          └──────────────┘
                                              │
                                              ▼
                                    ┌──────────────────┐
                                    │  smolified-      │
                                    │  krackhack26     │
                                    │  verilog (270M)  │
                                    └──────────────────┘

📄 License

This project uses the smolify/smolified-krackhack26verilog model. Please refer to the model card for licensing details.

🙏 Acknowledgments

• —Smolify AI — For the smolified Verilog model
• —KrackHack Hackathon — For the inspiration
• —HuggingFace — For the Transformers library