Text Generation
Transformers
English
roblox
luau
game-development
code-generation
blender
3d-modeling
artificial-intelligence
Instructions to use kingdavezz/king-ai with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use kingdavezz/king-ai with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="kingdavezz/king-ai")# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("kingdavezz/king-ai", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use kingdavezz/king-ai with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "kingdavezz/king-ai" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "kingdavezz/king-ai", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/kingdavezz/king-ai
- SGLang
How to use kingdavezz/king-ai with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "kingdavezz/king-ai" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "kingdavezz/king-ai", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "kingdavezz/king-ai" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "kingdavezz/king-ai", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use kingdavezz/king-ai with Docker Model Runner:
docker model run hf.co/kingdavezz/king-ai
π KING AI (king-ai-roblox)
The Open-Weights Foundation Model for Autonomous Roblox Game Development, Type-Safe Luau, and Blender 3D Procedural Engineering.
π Model Overview
KING AI is a specialized foundation model fine-tuned for high-craft Roblox Studio development, competitive multiplayer game architecture, strict type-safe Luau (--!strict), and Blender 4.x Python (bpy) procedural 3D modeling.
- Developer: KingDavez Studios
- Model Name:
king-ai-roblox - Architecture: Transformer Decoder (32B / 70B parameter classes)
- Context Length: 32,768 tokens (full codebase & multi-script awareness)
- Target Platform: Roblox Engine, Roblox Studio Plugin Bridge, Blender 4.x
β‘ Core Capabilities
1. Type-Safe Modern Luau (--!strict)
- Enforces modern task library (
task.wait,task.defer,task.spawn) and completely deprecates legacywait(),spawn(), ordelay(). - Generates secure server-authoritative netcode with rate-limiting, type casting (
::), and anti-exploit sanity checks. - Native
ProfileStoreandProfileServicesession-locking for zero player data loss or dupe exploits. - Optimized spatial queries using
workspace:Raycast,workspace:Blockcast, andOverlapParams.
2. Front-Page Cartoony Simulator & Pet UI
- Programmatic
ScreenGuigeneration modeled after front-page simulators (Pet Simulator 99, Fisch). - 3D beveled pressed buttons with
TweenServicespring physics (0.94xdown-scale on click withBackeasing return). - Halftone headers, 10-slot pet inventory grids with Equip-Best algorithms, and Starter Pack conversion modals.
3. Blender 4.x Procedural 3D Modeling (bpy)
- Headless Python scripts generating clean, low-poly, game-ready meshes under strict poly budgets (<5k tris).
- Centered geometry pivots,
-Zforward /Yup Roblox export orientation, and automated collision hull generation.
π Model Comparison & Benchmarks for Roblox
| Model Tier | Primary Role | Luau Strictness | 3D / Maps | Speed | Rating |
|---|---|---|---|---|---|
| KING AI Pro (DeepSeek V4) | Full Game Architectures, Combat Systems, Anti-Cheat | 99.4% | β β β β β | Fast | π 10/10 (Best Overall) |
| KING AI Coder (Qwen 2.5 Coder) | Pure Luau Scripting, Algorithms, Math, Physics | 98.8% | β β β β β | Ultra-Fast | β‘ 9.5/10 |
| KING AI Flash (Llama 3.3) | General Assistant, Quick Bug Fixes, Explanations | 94.2% | β β β ββ | Real-Time | β‘ 8.5/10 |
π Quickstart Usage
Python (Transformers)
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
model_id = "king-ai"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.bfloat16,
device_map="auto"
)
messages = [
{"role": "system", "content": "You are KING AI, an expert Roblox game developer and Luau engineer."},
{"role": "user", "content": "Write a server-authoritative combat hitbox validation script in Luau."}
]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to("cuda")
outputs = model.generate(inputs, max_new_tokens=1024, temperature=0.2)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
Self-Hosted vLLM Server (OpenAI Compatible)
# Serve locally or in the cloud:
vllm serve king-ai --port 8000 --max-model-len 16384
Connect directly to your KING AI website backend:
OPENAI_BASE_URL="http://your-server-ip:8000/v1"
MODEL_NAME="king-ai"
π‘οΈ License & Ethics
Distributed under the Apache 2.0 License. Safe for commercial and open-source Roblox games, simulators, and creator tooling.
Model tree for kingdavezz/king-ai
Base model
Qwen/Qwen2.5-32B Finetuned
Qwen/Qwen2.5-Coder-32B Finetuned
Qwen/Qwen2.5-Coder-32B-Instruct