Neural-Ink: Real-Time Stylized Universes

Vercel
application badge
Created by team VantEdge Neural Tech on May 07, 2026
AMD ROCmAMD Developer CloudAntigravityGemini AI
Vision & Multimodal AI

Neural-Ink V5 solves one of the most difficult challenges in real-time AI: bringing high-fidelity neural style transfer to the gaming industry without the common pitfalls of temporal flickering and high latency. Developed for the AMD Pervasive AI Hackathon, this project bridges the gap from Cloud to Edge. We utilize the massive compute power of AMD Instinct™ MI300X to train deep residual networks that process not just color, but 16-bit depth maps and motion vectors. The result is a temporally stable, 1930s-inspired "ink-style" aesthetic that runs at a lightning-fast ~15ms latency directly inside the Unity Game Engine. By leveraging ROCm 6.2 for cloud training and Unity Sentis (DirectML/Vulkan) for edge inference, VantEdge demonstrates how AMD hardware enables a new era of "AI-Native" graphics that empower creative teams to redefine visual storytelling.

Category tags:
GithubPresentationDemo
Explore more applications

Thymus

Thymus is a lightweight hybrid token-efficient router designed to maximize accuracy while minimizing token costs in multi‑task LLM pipelines. It dynamically routes user queries across local and remote models on LLM providers.

The Disappointer

HuggingFace HubLLaMAAMD Developer Cloud
Vercel
application badge

AI Classroom Edge Intelligence

A privacy-first classroom AI platform that routes sensitive work to local edge systems and eligible anonymized analysis to Fireworks AI, helping teachers make faster, safer instructional decisions even with unreliable internet.

AI Classroom Edge

AMD Developer CloudQwen3rest apiGithub CopilotCodexChatGPT

Taskly: Smart Multi-Model Task Router

Taskly classifies incoming tasks into 8 categories (QA, math, code, NLP) and routes each to the optimal Fireworks AI model with a tuned prompt — maximizing accuracy while minimizing token usage in a fully Dockerized pipeline.

RuntimeTerror

AMD Developer CloudAMD ROCm
Streamlit
application badge

router_007_v3

router_007_v2 is a Track 1 agent that records **zero billable tokens**: every answer is computed inside the container by a Qwen2.5-7B-Instruct model bundled in the Docker image and served in-process with llama-cpp-python

roc_auc_half

Claude CodeChatGPTAMD ROCmAMD Developer CloudCodexGemmaGPT-5NVIDIAQwen3

Lexyprep

Agent knows how to make thorough legal research based on my experience of winning legal cases. It understands your issue and advises on procedure

Lexyprep - Do you have a case

AMD Developer CloudAMD ROCmCodexGemma

"This is an impressive technical achievement that solves real problems in real-time AI graphics. Bringing neural style transfer to gaming without temporal flickering and high latency is a significant challenge. The 1930s-inspired ink-style aesthetic is creative, and the ~15ms latency is remarkable. The cloud-to-edge pipeline (ROCm training → Unity Sentis inference) demonstrates solid engineering. - New visual storytelling language for game developers - "AI-Native" graphics could be big in the future - Revolutionary for creative teams Application of Technology: 🚀🚀🚀🚀🚀 5 - Uses AMD MI300X for training deep residual networks that process 16-bit depth maps and motion vectors (not just color). Unity Sentis (DirectML/Vulkan) for edge inference. ROCm 6.2 for cloud training. ~15ms latency is excellent. Presentation: 🚀🚀🚀🚀🚀 5 - Excellent problem statement (temporal flickering, latency). Clear technical explanation. Demo link provided. Well-structured with good visuals. Business Value: 🚀🚀🚀🚀🚀 5 - Gaming industry is massive. Could empower creative teams to redefine visual storytelling. "AI-Native" graphics is an emerging field with huge potential. Originality: 🚀🚀🚀🚀🚀 5 - Very original. The 1930s-inspired ink-style aesthetic is unique. Processing depth maps and motion vectors is innovative. Addresses temporal stability which is a known problem in neural style transfer. "

avatar

Sanem Avcil