Pravahini (प्रवाहिनी) - Where Data Flows, Models Grow, and AI Glows!

Hi @victorious, here is a very crucial bug which I think you need to clarify.

The idea of implementing marketplace for people to sell model and datasets using Tron token assets. However, Hugging Face being the biggest ml community, they have so many models and datasets (which you can think about) and it is free to use. What makes you stand out in this awkward situation since your project concept is Web3 version of Hugging Face.

2 Likes

Currently, since people have such a big repository of state of art model and complete datasets, maybe a Web3 version Hugging Face which just paying with Tron token assets will chase people away. How do you think?

1 Like

Plus, AI track wish to promote projects which pushes AI innovation on Tron network. If your project just focusing on marketplace to buy and sell model and datasets, they can just simply do this action in a Web2 open source free repositories and use it to build something on Tron, which can easily skip your project in this process

Maybe you can design some mechanism to actually stand out from this

Hi @tanzhixuan!

Thank you for your thoughtful questions and your perspective. :smiley:

Our project involves creating a platform that leverages Web3 technologies, particularly the TRON/BTTC blockchain, to promote AI innovation and contribute to an AI-centric future which is mentioned in the theme of AI Track.

We understand the abundance of state-of-the-art models and datasets available in the current landscape. While our project shares some similarities with platforms like Hugging Face but it’s not exactly the same, we’re excited to highlight you key aspects that set us apart and bring added value to the table:

  • Monetize Opportunities: Unlike platforms offering entirely free content, we empower AI Engineers to monetize their hard work. By allowing users to set prices for their datasets and models, we create a marketplace where quality and value are rewarded.

  • User-Controlled Access: With different visibility options (public, private, for sale), our platform allows users greater control over whether they want to keep within themselves, or share with everyone or sell that particular model or dataset. This can be valuable for creators who want to carefully manage the distribution of their work.

  • Decentralized and Blockchain Elements: If we incorporate blockchain technology, our platform might offer enhanced security, transparency, and trust for transactions, and ownership.

  • Private Content Sharing: Our platform supports private content sharing, allowing users to keep their datasets and models private. This can be appealing to those who need to protect sensitive or proprietary information.

  • Decentralized Computation: With this feature Users with limited hardware resources, such as low-spec laptops, can access high-performance computing power for running resource-intensive tasks. They can harness the power of distributed processing by submitting GitHub URLs for ML models and datasets. This approach offers efficiency and scalability for data processing tasks. This aligns with the theme of harnessing blockchain technology to enhance AI capabilities.

  • Coding Interface Integration: Our all-inclusive platform includes a coding interface, enabling users to write and execute code seamlessly.

  • Add on Features: In addition to the marketplace, we are planning to offer plagiarism detection and community driven governance. These functionalities enhance the AI innovation process beyond simple model and dataset trading.

Our project encourages the implementation of AI solutions using TRON/BTTC blockchain, which directly aligns with the hackathon’s goal of utilizing blockchain for AI-centric solutions.

It boasts a unique design and user experience, offering a refreshing and innovative environment for dataset and model sharing.

All the features contribute to a positive impact by empowering engineers and researchers to collaborate, innovate, and share within a decentralized AI ecosystem.

So we believe that this combination of features makes our project different from them. We appreciate your feedback and will continue to enhance our project in alignment with the AI track’s goals. If you have any suggestions or any questions please feel free to ask. :grinning: :raised_hands:

1 Like

Thanks for your answer, the ideas are quite promising. But is all the listed features will be introduced in this hackathon? I am curious about the implementation issue in such a short timeframe, what is your plan on this? As instead of marketplace, you even need to implement live computation and also a github space concept coding environment on your platform

2 Likes

Hi @tanzhixuan !

Thank you for showing your excitement and interest in our project. :grinning:

Our project is moving ahead as per our planned sprints. We’ve successfully completed Sprint - 1 and are currently Sprint - 2 is in progress. Frontend development is near completion, along with front-end, implementation of Smart Contracts is in progress.

We’re also making good progress with the creation of APIs for Decentralized Computation, to run the computation jobs.

We are trying our best to deliver all the functionalities within the designated time frame. Once again Thank You and if you have any questions please feel free to ask. :raised_hands:

1 Like

Hi @Ekuns !

Thank You. Welcome to the Tron Hackathon S5! :rocket:

1 Like

Hi @manfred_jr !

Thank You for the warm welcome and sorry for replying late.

Our platform utilizes Bacalhau, for decentralized computation. When a user submits a Raw GitHub URL of an ML model (Public Repo), Bacalhau fetches the model and performs distributed processing using WebAssembly (wasm) images and Docker containers.

This approach optimizes the computation process by leveraging the combined power of the network while ensuring reliability and security.

If you any additional questions feel free to ask. :grinning: :raised_hands:

Thanks for replying, I understand running a project and keeping up with project submissions during hackathon can be a very tedious task.
From your reply I easily understand that incorporating Bacalhau for decentralized computation is a significant enhancement to the Pravahini platform. Leveraging Bacalhau’s capabilities to process ML models using WebAssembly (wasm) images and Docker containers from Raw GitHub URLs of public ML model repositories can bring several benefits to the project.

Are you open to suggestions though? If yes, here’s one if I may; given the complex nature of distributed computation, providing an intuitive user interface that informs users about the status of their tasks, progress, and successful completion could enhance user satisfaction.

Moving forward, I have just one question;
How was Bacalhau selected as the decentralized computation solution? Were there any specific features or benefits that led to its choice over other alternatives?

Gm and a happy weekend to you
Adding a happy emoji to the last part of your reply lit up my mood, thanks!

1 Like

please can your platform serve the same purpose like Chatgpt?

Thank you @manfred_jr for understanding. Yes we are always open for the suggestions. :raised_hands:

So we have tried to give the instructions in the computation page and also as you mentioned about the tasks, progress and completion of that task, we have kept that. So user can easily understand what’s going on and it will be easy for them to use.

For computation we selected Bacalhau because their documentation is quite good and easy to understand and if we are stuck somewhere then the community members are supportive in replying and giving feasible solution. And the features provided by the Bacalhau such as execute computations on decentralized network, manage jobs, provide verifiable results and the performance are the factors that align with our project’s needs.

If you have any questions or any suggesstions feel free to share. Have a great day! :grinning:

5 Likes

Thanks for your question @Prince-Onscolo . I am not sure how our platform can serve like chatGPT. We focus more on decentralized computation and a marketplace for datasets and models.

However, if you have any suggestions regarding how it can serve as chatGPT we can surely keep that in mind for the future.

oh OK thanks for the response

1 Like

Thanks for replying with some additional insights and information. I’m wondering;

Could you provide more details on how Bacalhau’s performance aligns with the project’s needs? Are there any specific benchmarks or metrics that have been evaluated to confirm this alignment?

And I’d like to make some suggestions, which of course as far as all my suggestions goes it’s always applicable generally;

Explore ways to integrate the supportive Bacalhau community directly into the platform. This could include a dedicated section for discussions, FAQs, or even integration with relevant Bacalhau forums.

Clearly communicate to users how the verification process works for computation results. Transparency regarding data integrity and authenticity can build trust.

Create onboarding resources for new users, including step-by-step guides, video tutorials, or interactive demos to help them get started with the platform and Bacalhau seamlessly.

4 Likes

Love how detailed your planning is!

This gives a good opportunity for small AI model creators to find buyers and monetize their work.

Wish you all the best for the hackathon :raised_hands:

1 Like

Sure @manfred_jr ,

These are some of the important points that how Bacalhau aligns with the project needs:-

Scalability: Bacalhau’s architecture allows it to scale efficiently, which is essential for handling a large number of computation jobs. This aligns with the project’s need to process a high volume of computations effectively.

Execution Speed: Bacalhau can execute computational tasks quickly.

Security: Bacalhau provides a secure execution environment for computations. It isolates jobs, which helps protect sensitive data and code. This aligns with the project’s need for secure computation handling.

Decentralization: Bacalhau is a decentralized computation platform, which means it can utilize idle resources from various nodes in a network.

And thank you for providing the valuable suggestions. We will surely give it a thought. We are up for all the innovative suggestions! Have a good day! :smiley:

2 Likes

Thank you for your kind words and well-wishes @Nyx !

We appreciate your support and enthusiasm for the project. The aim is indeed to create a platform that benefits both AI model creators, Dataset Creators and those in need of AI services.

If you have any more questions or need further information in the future, feel free to reach out. Have a great day! :rocket::+1:

2 Likes

Wishing you success as hackathon progresses, would be keeping a close watch as things unfolds eventually, ciao!

1 Like

Please i want to know how AI model creators will monetize their creation on your platform.

How will the flow be like?
Will be very happy if it can be explained a little

1 Like

Yes sure @manfred_jr ! Thank for your wishes. :raised_hands:

1 Like