Zero-to-Hero Setup Guide

Mandatory Pre-Workshop Preparation | Follow these steps exactly.

🏁 Goal: Arrive with a fully offline-capable AI PC.
⚠️ Warning: Total download size is ~20GB. Do NOT do this on mobile data or at the venue.

🚨 Having Issues? Verify your 7zr.exe and Model Paths.
Open Troubleshooting Guide

Phase 1: System Foundation

1. Update Intel Drivers

Ensure your NPU and GPU are recognized correctly.

Download Intel Driver Assistant

Run the assistant -> Install all "Graphics" and "NPU" updates -> Restart PC.

2. Install Python 3.10

Specific version required for compatibility.

Download Python 3.10 Installer

CRITICAL STEP: During installation, check the box "Add Python 3.10 to PATH" at the bottom of the first screen.

3. Install VS Code

We will use this to edit configuration files.

Download VS Code

Phase 2: AI Playground (No Code)

1. Install AI Playground

The main interface for Chat and Image generation.

Download Latest Release (GitHub)

Look for the .exe installer in "Assets". Run and install.

2. Download Models (In-App)

Open AI Playground application. Go to "Models" or "Manage" tab.

LLM: Download Phi-3-mini-4k-instruct (approx 4GB)
Image: Download SDXL Turbo (approx 6GB)

Tip: Keep the app open until download bars reach 100%.

Phase 3: Creative Tools

1. Install GIMP

The free image editor we will supercharge with AI.

Download GIMP

2. Install OpenVINO Plugins

Connects GIMP to your Intel NPU.

Download OpenVINO GIMP Installer

Download openvino-ai-plugins-gimp-setup.exe. Run it. It will auto-detect GIMP.

3. Initialize Plugins

1. Open GIMP.
2. Click Layer in top menu -> OpenVINO-AI-Plugins.
3. Select Stable Diffusion.
4. Wait for it to download models (This takes time!).

Phase 4: Developer Environment

1. Install Ollama

The local server for running LLMs via API.

Download Ollama for Windows

2. Pull Ollama Models

Open your Command Prompt (cmd) or PowerShell and run these exact commands one by one:

ollama pull llama3
                    ollama pull llava
                    ollama pull nomic-embed-text

Wait for each to finish (Success status) before running the next.

3. Install Python Libraries

We need these for the Voice AI and Agent demos.

pip install --upgrade pip
                    pip install pipecat-ai
                    pip install aiohttp
                    pip install zerodha_kit
                    pip install colorama

Phase 5: Final Verification (The "Offline Test")

Pro Tip: Disconnect your Wi-Fi now to perform these checks. If they work offline, you are ready!

Check 1: AI Playground

Open App -> Select "Phi-3" -> Type "Hello, who are you?".
Success: It replies quickly without internet.

Check 2: GIMP

Open GIMP -> Layer -> OpenVINO -> Stable Diffusion.
Success: The menu opens and you can click "Generate".

Check 3: Ollama

Open Terminal -> Type ollama run llama3.
Success: You get a prompt >>> and can chat.

Need help? Contact the workshop coordinator at [Your Email/Discord]