hsosa/gaming-ai-corpus

🎮 Gaming AI Corpus & Dataset Community

A massive, structured collection of multi-game wiki text and scripting data designed for training, fine-tuning Large Language Models (LLMs), building Retrieval-Augmented Generation (RAG) systems, and developing custom video game AI agents.

---

🌟 Support & Project Resources

If you find this dataset useful for your AI or RAG projects, please support the project by leaving a star on GitHub!

• GitHub Repository: huso3123/gaming-ai-corpus ⭐
• Implementation Video Guide: Natures-g8u on YouTube 📺

---

📊 Dataset Overview & Features

This dataset contains raw, cleaned text corpora scraped and parsed from major game wikis, formatted cleanly for natural language processing tasks, model fine-tuning, and embedding generations.

• Universal Compatibility: Compatible with any LLM architecture (including Llama, DeepSeek, Qwen, Mistral, GPT) for custom fine-tuning (LoRA/QLoRA) or RAG injection.
• Massive Volume: Contains over 70,000+ detailed lines per main category, adding up to a rich text corpus optimized for tokenizers.
• Deep Knowledge Base: Offers highly detailed game universe mechanics, item descriptions, and comprehensive code logic templates rather than just shallow summaries.

Included Files & Structure:

• Minecraft Corpus (mchakkindakihersey_turbo.txt) - Comprehensive Minecraft wiki guides, items, mechanics, and lore data (70,000+ lines).
• Roblox Corpus (roblox_70k.txt) - Roblox Luau mechanics, scripting guides, game systems, and wiki knowledge base (70,000+ lines).
• GTA V & FiveM Corpus (gta_v_game_dataset.txt) - Grand Theft Auto V game mechanics, mission structures, dialogue data, and FiveM scripting documentation (LUA / native functions).

---

🚀 Quick Usage (Python)

You can easily load and use these text files in your Python machine learning or RAG pipeline:

# Example of loading the Minecraft, Roblox & GTA V corpora
with open("mchakkindakihersey_turbo.txt", "r", encoding="utf-8") as f:
    minecraft_data = f.read()

with open("roblox_70k.txt", "r", encoding="utf-8") as f:
    roblox_data = f.read()

with open("gta_v_game_dataset.txt", "r", encoding="utf-8") as f:
    gta_data = f.read()

total_chars = len(minecraft_data) + len(roblox_data) + len(gta_data)
print(f"Loaded {total_chars} characters of gaming knowledge across 3 universes!")

---