Image generated with Stable Diffusion v0.9 run on Lilypad: lilypad run sdxl-pipeline:v0.9-base-lilypad3 -i Prompt="digital lilypads in a cosmic library turquoise tones" -i Steps=200
🌎 Overview
The Lilypad Network is a serverless, distributed compute network providing internet-scale data processing for AI and other next-gen use cases.
Lilypad offers a coordination platform for compute resources. The network brings together the exponentially rising user demand for compute services and hardware with compute providers who can add new revenue streams and optimize efficiency to their hardware business.
By leveraging peer to peer networking and crypto-economic incentives, the Lilypad Network performs off-chain decentralized compute over data, with on-chain verification using a distributed network of GPU nodes.
To run AI Inference and other data processing jobs, Lilypad provides a range of tooling including a CLI, smart contracts, and API. The network has been built to serve use cases, including Inference AI jobs, ML training jobs, invoking & supporting generic ZK computations, leveraging inbuilt storage on IPFS, Federated Learning consensus (with Bacalhau insulated jobs), providing a platform for Digital twins, and more!
The Lilypad Incentivized Testnet has been deployed to Arbitrum Sepolia and will launch publicly in mid June! As the IncentiveNet launches, our team is proud to on-board GPUs from key partners including Swanchain, Spheron, and Rare Compute!
Running from June 2024 through to the launch of Lilypad on mainnet (anticipated in Q4 2024), IncentiveNet presents a unique opportunity to scale and test core compute infrastructure critical for the viability of the Lilypad economy. Participants will be rewarded in the form of Lilybit_ rewards, which will be redeemable for the Lilypad ERC20 Token at Mainnet Launch.
The Lilypad team will be at ETHCC in Brussels, Belgium!! If you will be in town, reach out to our team on Discord. We’ll post a list of events we are attending soon.
Sign up to the Lilypad Newsletter or follow the Lilypad Network on X.
⚒️ Engineering Update
We're thrilled to announce the Lilypad Network has transitioned to theArbitrum Sepolia Testnet, preparing for the full migration from the current Geth-based Testnet! This was a critical step to launch the Lilypad IncentiveNet and will be a foundational L2 for mainnet launch.
The Lilypad Resource Provider Leaderboard will launch soon (after some minor improvements)! This will be the central hub for Resource Providers to track Lilybit_ rewards earned during the Incentivized Testnet Program. The leaderboard ranks the top Resource Providers by calculating their “Energy Provided” (TFLOPS*s).
Lilypad deployment to the Arbitrum Sepolia Testnet
Lilypad orchestrates job processing through Job Creator and Resource Provider (RP) nodes, utilizing Arbitrum L2 for off-chain compute verification and validation. This shift enhances the scalability of the Lilypad Network by significantly lowering transaction costs and increasing throughput, all while maintaining compatibility with Ethereum-based tools and DApps.
Leveraging a variety of CI/CD workflows, our team was able to build, test, and deploy the Lilypad Testnet to the Arbitrum Sepolia Testnet within an extremely short window of time! In addition, metrics and logging mechanisms continue to be improved to ensure network reliability and public observability for the community.
With the move to Arbitrum, the Lilypad Testnet faucet will be moved over to support the Arbitrum Sepolia Testnet. We will keep the community updated as this change is completed!
Resource Provider Leaderboard
The Resource Provider Leaderboard and Metrics dashboards will serve as public tools for the community to monitor progress of the network. These tools provide an overview of RP capacity on the network as well as points earned by each address!
We believe it’s important to build in public and are proud to prioritize transparency on important network metrics. A Resource Provider’s power is calculated by the total hashrate reported to the network. Each RP will run a PoW calculation that will allow the network to measure the difference in compute between, for example, a Nvidia 3090 and 4090ti.
Stay connected on the Lilypad Discord and follow along on X/Twitter for the latest on the community leaderboard and new website!!
Module Allowlist
To improve safety and security, a public Lilypad Module allow list has been implemented. This mechanism helps the network avoid running malicious programs added as Lilypad modules. It’s critical that this process of adding modules to the allowlist be transparent and inclusive of the community. Resource Providers can be assured that their hardware is not processing illegal or dangerous programs.
The Lilypad team and community is able to block modules network wide and Resource Providers are able to specify what modules run within their Resource Provider nodes as well.
Feel free to reach out to us in Discord and help us decide on modules to add. In addition, a community voting mechanism will be rolled out in IncentiveNet Phase 2.
🖥️ Tokenomics & IncentiveNet
Phases 0 and 1, aim to incentivize reliable GPUs to join the Lilypad IncentiveNet. Lilybit_ rewards will be awarded for time on the network (up to a 4x multiplier) and compute power brought to the network. The earlier a provider joins the network, the more Lilybits_ will be available. 5-10% of the total supply of Lilypad Tokens at mainnet launch will be allocated to RPs and Developers with Points.
A Proof of Work mechanism has been implemented for Resource Providers to report a hash rate to the network. More details will be released on this soon!
Incentivised Testnet Program - Lilybit_s
The IncentiveNet will run continuously from June 2024 through to the official launch of Lilypad on mainnet - anticipated in Q4 2024. Lilybit__ reward points will be redeemable, for the Lilypad ERC20 Utility Token at Mainnet Launch, with between 5% and 10% of the total token supply (depending on IncentiveNet demand and tokenomics finalization) being allocated to this phase of the Lilypad Network.
Earning Lilybit_rewards is directly relatable to useful work being performed on the network or on behalf of the project goals. These goals are also holistic in nature to encompass the need for business validation and community education initiatives while building feedback loops with users to improve the network.
The goals for the Lilypad IncentiveNet fall into the following main categories:
Network Growth Goals (→ supply, jobs, and ecosystem)
Network Security & Readiness Goals (Testing in Real-world conditions)
Community Building, Engagement, and Education Goals
Market Readiness, Research, Product Testing & Business Validation Goals
The overall distribution of the Lilybit_ reward credits will be heavily skewed in favor of those providing essential services to the network growth - GPU compute providers, with a rough heuristic of 70% of rewards directed towards network growth goals and 30% towards all other activities. Lilypad aims to recognise the contribution of all types of community members, including non-technical participants who will also be able to earn rewards for their value-added activities in the Lilypad Network. Reach out to our team with questions in Discord.
🔬 Verification and Privacy Research
Our team has conducted an in-depth analysis of a potential strategic partnership between Lilypad and Zama (https://www.zama.ai/), revealing that a collaborative effort could effectively harness their complementary strengths to drive innovation and growth for both entities.
Data privacy using mechanism like Homomorphic Encryption (FHE) open up new possibilities for end user privacy as well as business use cases. We are excited about this future and are taking the time to understand how our network can leverage this revolutionary technology.
Research is ongoing with regards to emerging demand and offering segments, each with distinct needs and requirements. Various technologies are being tailored to meet the unique demands of disparate customer groups.
⚡ Developer Events & DX Updates
DevNTell AMA
Ally Haire and Phil Billingsby joined the latest DeveloperDAO DevNTell event to talk about all things Lilypad and demo running jobs on the network! We greatly appreciate the DeveloperDAO ecosystem and will continue to support the community. If you missed the event, check out a recording below!
Weekly AMA Sessions!
To help answer questions and gain insights from our team, Lilypad is hosting AMA sessions every Wednesday! We will alternate the timing of the AMA every week to ensure we can cater to RPs and devs across the globe! Find an AMA time that works for you on the Lilypad Calendar (https://lu.ma/lilypadnetwork).
🚀 Partnerships
Lilypad partnership with Rare Compute!
We're thrilled to announce a partnership between Rare Compute and Lilypad, enhancing our GPU capabilities. The Rare Compute Project and Lilypad are collaborating to unlock the power of GPUs into the Lilypad Network, significantly enhancing our computational capabilities.
This initiative allows for providing GPU power via containers directly to users, opening up opportunities for advanced scientific research and AI applications. The project supports various use cases by enabling GPU rental on an hourly or credit basis, including decentralized scientific (DeSci) computing and rare disease research. This collaboration aims to tell impactful stories and demonstrate the critical role of decentralized computing in scientific advancements.
Join us on Discord to witness the real-time development of this exciting project!
What's on the Horizon?
While much partnership work is still in process, we are prioritizing a few key integrations to improve the Lilypad experience:
- On-Chain Databasing & Storage:
We are thrilled to work with one of the leading distributed database providers leveraging Filecoin for storage and IPC for scaling data management needs. This tech stack provides enhanced security, transparency, and trust for users of the Lilypad Network. Stay tuned for more on this partnership.
- Novel Web3 Applications (dApps):
Our DevRel team is focused on a world class developer experience. This involves creating new tools and experiences which wouldn't be possible without a decentralized compute infrastructure.
Our research team has prioritized work on understanding the best privacy preserving mechanisms to implement into Lilypad. Innovations such as Homomorphic Encryption and Trusted Execution Environments could provide avenues to accomplish our privacy goals!
HackFS bounty winners announced!
With 15 total project submissions, the HackFS projects built on Lilypad blew us away! Teams focused on end-user applications while leveraging Lilypad infra and CLI to spin up their projects quickly. Teams built AI agents, AI model marketplaces using Filecoin, and AI inference with encrypted results using Lit protocol. Read more about the event below!
1st Place - Best implementation of Lilypad
CypherCraft aims to simplify the process of training models on private datasets. These datasets can have custom access control conditions that enable specific types of users or model providers to use these private datasets. CypherCraft automates deploying Machine Learning (ML) model scripts by auto-containering, training the models on the Lilypad compute network, and saving the models weights on IPFS nodes.
2nd Place
Nextown created a marketplace for running AI models. First, users subscribe to the Nextown channel to run preexisting AI models or add new models. Then, they select AI models from the UI, which are executed on the LilyPad decentralized compute network. The results are then encrypted using the user’s public key to ensure privacy via Lit Protocol and stored on the lighthouse.
3rd Place
Filecoin ML Engine is an onchain ML agent built on Lilypad to provide data collection, extraction, and computation on top of the Filecoin virtual machine.
4th Place
Dog token AI - A decentralized application built on the ERC-404 standard. When users hold the project's tokens, they are also automatically granted a unique, non-fungible token (NFT). The NFT doesn't have a pre-generated image. Unleash your creativity and bring it to life using Lilypad AI's generative art technology.
🔮 What's Next?
As the Incentivized Testnet launches, we are committed to ensuring RPs have a smooth on-boarding process. This includes metrics reporting, consistent communication on system upgrades, up to date docs and guides, and more.
Each of the Lilypad Incentivized Testnet phases has a different goal. Phase 1 is focused on incentivizing GPUs to join the network. This will take place over the month as Resource Providers join the network. The Lilybit_ rewards program will kick off during Phase1 and will carry on through to mainnet launch.
The next phases of the IncentiveNet will continue to reward Resource Providers as well as provide incentives for job modules added to run on Lilypad, plugins/integrations, and more! Stay tuned for more on this.
💬 Have more questions or looking for guidance? Join our community AMA hosted every Wednesday!
Chat to us on Discord and follow along on X/Twitter for the latest news and updates!
