News Board

• [PIN] This project is part of the PALISADE-X project funded by DOE ASCR.

• [PIN] We preprinted a paper describing the design of APPFLx on arXiv.

• We have integrated HuggingFace with APPFLx! You can now employ pretrained model on HuggingFace for federated learning experiments on APPFLx.

• We added a resource monitor in the Federation Information page for checking the resource utilizations of the client computing endpoints.

• We supported Federated Asynchronous algorithm in APPFLx for asynchronous federated learning experiments!

Roadmap

• Client data distribution visualization: We will soon add the capability for users to check the distribution of data accross different federated learning sites.

• FedBuffer asynchronous algorithm: Integrate FedBuffer into APPFLx.

• LoRA fine-tuning: Integrate LoRA fine-tuning methods to APPFLx to support fine-tuning LLM in a federated manner.