Jeton.AI
Edge AI Computing for Web3
Large AI Model Acceleration Platform on Edge Device for Web3 World
Jeton.AI: Running LLM Effectively
on Mobile Device
Fine tunning/domain tunning/QAT model to fit specific AI needs with best intelligence
SOTA technique to compress the LLM model
Leading algorithm to maximize the computing power of a Mobile Device
Multimodal/Hybrid mode support of Text Chatting, Visual contents generation and Understanding
Jeton.AI Unique Strength
- •Personalized/Compressed AI Model on mobile device
- •Unique Edge Computing Accelerator for Large Model running on Mobile Device
- •LLM and Visual AI Acceleration
- •Open SDK to build embedded AI DApp
- •Open Source of the Edge Computing Framework
- •World's Top AI Performance Optimization Team
Jeton.AI for On-device Multi/hybrid Model LLM AI
Jeton.AI: For On-device generative AI:
QWen2: 1.5B
Gemma2: 2B
LLAMA2: 7B
ChatGLM: 6B
Mistral-7B-Instruct: 7B
SD Model
Multimodal Model etc.
with native Huggingface compatible
With SOTA int4bit/3bit model quantization
Jeton.AI Can be Easily Integrated with Web3
Jeton.AI Unified Platform
Text-to-Text AI
- Intelligent Chatting Bot
- Intelligent Reasoning/Agent/RAG
- Can be easily integrated
- Easy to Use
- Crypto Tokens for LLM Tokens
Text-to-Image AI
- Support Visual Content Generation
- Support native Diffusers
- Graph extract and Optimization
- Graph Fusion
- Visual Contents Understanding
Hybrid/Multimodal
- Integrated UI for both Text and Visual Contents
- Smooth Transition
- Easy to Use
No GPU required
Data Sharing can earn Crypto Tokens
OpenAPI for Partner/DApp Integration
Saving AI Infra Cost for up to 50% for DApps that provide AI Service
- • Model Pruning Service Edge AI Node
- • Acquiring Service Hybrid Edge
- • Acceleration with Cloud AI Service
Saving AI Infra Cost for up to 50% for DApps that provide AI Service
- • Model Pruning Service Edge AI Node
- • Acquiring Service Hybrid Edge
- • Acceleration with Cloud AI Service
Orbit Coin Tokenomics (Draft)
Tokens total:
• TBD
Token Inflation:
• Next 2 years 20% (Compensate for the growth of supply and demand, predicted growth rate 60%YOY)
Tokens distribution:
- • xx% the Team
- • xx% the Communities (10% reserved for the first year consumption, 50% reserved)
- • xx% for the rest
Tokens mining (Model Data Provider) determined by the following factors:
- • Developing App for with AI Edge SDK
- • Time of App Usage
- • Customized Model Data sharing
- • Idling time
Token burning (Accelerator Users) determined by the following factors:
- • LLM Tokens (exchanged to Crypto AI Tokens) used
- • Required Accelerator types
- • 5% Burned
Tokens total:
• TBD
Token Inflation:
• Next 2 years 20% (Compensate for
the growth of supply and demand,
predicted growth rate 60%YOY)
Tokens distribution:
- • xx% the Team
- • xx% the Communities (10% reserved for
the first year consumption, 50% reserved) - • xx% for the rest
Tokens mining (Model Data Provider)
determined by the following factors:
- • Developing App for with AI Edge SDK
- • Time of App Usage
- • Customized Model Data sharing
- • Idling time
Token burning (Accelerator Users)
determined by the following factors:
- • LLM Tokens (exchanged to Crypto AI Tokens) used
- • Required Accelerator types
- • 5% Burned
World Best AI Performance Optimization Team
- • Expert of Best Workload Agnostic AI Acceleration Engine
- • Leading the Development of MNN.zone of Edge AI Computing
- • Expert of Large Scale AI/Rendering Workload Optimization
- • Leading Expert in GPU/FPGA/NPU Virtualization and Sharing
Breaking AI Performance Records:
- • Stanford DAWN Bench AI Benchmark World Record Breaker
- • TPCx-BB Benchmark World Record Breaker
APAC Largest Public Cloud GPU/FPGA/NPU Service's Technical & Business Founder of A* Cloud
Academic Contributions
- • Top AI/Accelerator Conference Publications
- • GPU/FPGA/NPU Patents worldwide
International Teams located in US, Singapore, HK