Spaces:
Sleeping
Apply for community grant: Academic project (gpu and storage)
As part of the Artificial Intelligence for Philanthropy initiative at University of Geneva, Bern University of Applied Sciences and Philea, we are building an AI powered LLM for philanthropy enabling organisations to find others tackling similar societal or environmental challenges and this way foster meaningful exchanges, increase the sector visibility, and promote the formation of partnerships. We want this space to host an LLM able to learn from the inputs and examples from philanthropists for the RAG task in our academic database of philanthropic organisations and missions, therefore we need like to use the FlagICLModel or a similar one. The API for this model does not allow ICL capabilities, therefore does not serve our purpose. We opted for downloading and using the model at runtime, but with CPU it takes up to 5 minutes to perform a RAG search : ( , we need GPU if we want this tool to be usable so philanthropic organisations see what is possible with a tool like this available for the community. Unfortunately we do not have funding still to carry out this project with a budget, so currently is an altruistic initiative with hopes that will be accepted and supported by the community once out there. We would greatly appreciate your support for a GPU free or discounted tier to show the tool to the philanthropic community so we can get their support in the long term and keep the platform at huggingface. The persistent storage is also desirable since we are now incorporating a SERPapi websearch functionality for philanthropists to be able to search, download and query annual reports. If the space has our academic database and the reports users find, and keeps on growing, the space will get smarter and smarter with time. Thanks in advance for your support, we hope you find this project as important as we do.