search

vaseline555 / SuPerFed

(SIGKDD 2022) Connected Low-Loss Subspace Learning for a Personalization in Federated Learning (https://arxiv.org/abs/2109.07628)
28 stars 4 forks source link

About shakespeare dataset #2

Closed MirandaHe closed 1 year ago

MirandaHe commented 1 year ago

Thank you for your outstanding work. When I tried to run shakespeare's data set, I encountered the following problems. Do I need to download the data set by myself and construct the required directory path?

image

Looking forward to your reply

vaseline555 commented 1 year ago

Hi, thank you for your interest in my work!

Yes, you should prepare Shakespeare dataset manually, according to the instruction described in the original LEAF benchmark repository. Please see: https://github.com/TalwalkarLab/leaf/tree/master/data/shakespeare

Since it is somewhat bothersome, I've recently released another repo (for FL research) that supports automatic preparation of all datasets in LEAF benchmark. Please see: https://github.com/vaseline555/Federated-Learning-PyTorch

BTW, I have not yet completed transplanting SuPerFed into that repo. I hope you find this information helpful but if you require any further clarification please let me know.

Thank you.

Best, Adam

MirandaHe commented 1 year ago

Thank you for your answer, which solves some of my confusion. Ok! I'll learn about your new job. Thank you very much! Please forgive me if I'm new to this field。

vaseline555 commented 1 year ago

No problem at all, and welcome to FL community. 😄 Fingers crossed!