A blockchain-based protocol for trustless evaluation and purchase of machine learning models on Ethereum.
DanKu is a blockchain-based protocol that creates a decentralized marketplace for machine learning models on Ethereum. It allows users to post datasets and monetary rewards, inviting participants to submit trained neural networks, which are then evaluated on-chain to determine the best model and facilitate payment.
Machine learning practitioners, researchers, and organizations looking to monetize AI skills or crowdsource ML solutions in a trustless, decentralized environment.
It eliminates intermediaries by using smart contracts for objective model evaluation and payment, providing a transparent, global platform that incentivizes high-quality AI development and broadens access to machine learning.
Exchange ML models in a trustless manner!
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Uses Ethereum smart contracts to create a trustless marketplace, eliminating intermediaries and ensuring transparent transactions, as outlined in the protocol's whitepaper.
Executes submitted neural networks directly on the blockchain against posted datasets and evaluation functions, guaranteeing fair and unbiased assessment without human intervention.
Provides ML practitioners a direct way to earn rewards by submitting trained models, bypassing traditional platform fees and enabling global participation.
Enables anyone worldwide to post datasets or submit models, fostering innovation and broad crowdsourcing of machine learning solutions.
Requires installation of Solidity compiler, virtual environment, and Populus framework with multiple dependencies, making initial development complex and time-consuming.
On-chain execution of neural networks on Ethereum incurs significant gas fees and slow processing times, rendering it impractical for cost-sensitive or time-critical applications.
Datasets must be posted publicly on the blockchain for evaluation, compromising data privacy and limiting use cases involving confidential or proprietary information.