@Kasper_sky  The promotion of this project is planned as posts in Twitter and Discord communities. It will help to collect feedback from users and correct future development track as well as introduce it to the wide audience. The entire project is oriented to the new members of the community to help them figure out the mechanics of the governance and make their decision simpler. But in the future it can grow by the suggestions from experienced users to make it helpful even for them. So, the main goal of the currently planned promotion is to make it easy to find info about this tool. And I hope it will lead to the foundation of an entire section dedicated to the development of tools for lowering the entry threshold in the governance infrastructure to make more people participate in the referendum.

@Adam_Clay_Steeber  This is a request for compensation of already done work as well as for realization of future plans with mentioned above aims. After this stage I’m planning to have feedback from community members on hand and make a plan for further development. My tech stack is based on msmarco-bert-base-dot-v5 model for general data processing, punkt model for sentence analysis, polkassembly api for data collection and general purpose python libs for the rest of processing. Yes this is planned as an open source project for everyone to be able to clone it and run locally with your modifications and make merge requests. There is a git implemented in Huggingface so there is no dedicated Github repo. For now I’m alone but there is some possibility of expanding to team work. For the current stage of this project Huggingface is the optimal solution as it’s capable of handling several thousands requests per hour and easy to operate. Of course, when the project grows it will require more mature hosting. The probability calculation requires a lot of data to be processed. So for acceptable accuracy there should be at least 10 times more proposals in the database. And therefore realization of this function directly depends on community activity. And thank you for your feedback about proposal content, I'll take it into account next time

@Daniel 

I've decided to vote NAY after this comment for the following reasons:

• You will almost certainly never get enough data to train any meaningful model. You said you need 10x more data? That'll literally take years at this rate.
• No dedicated github and you're alone. This project needs a team of at least 3 people and proper branding. Otherwise this will become a tool that no one uses and isn't supported.
• No mature proposal. I didn't ask for a more mature proposal next time. I wanted one this time. You need to formalize this project so it doesn't become aimless and obscure.

You clearly have the developing capabilities to make this work. but this project completely lacks cohesion. Please go back to the drawing board and ask yourself how this project can become a highly useful tool for years to come that people are aware of.

@StakerSpace  Hi, Well, as I’ve mentioned in my reply, the code is open sourced and you can find it on Huggingface https://huggingface.co/spaces/DanichOne/Proposal_Similarity_Kusama/tree/main. Basically the dedicated website right now doesn’t give any advantages over the Huggingface as it’s just a way of hosting and the actual interface is Gradio. I’ll try to host it on Vercel as it seems to be a good starting point but it will take some time to switch to another platform. It will be published in the community social networks as well as Huggingface alternative.