med-air / fedbn Goto Github PK
View Code? Open in Web Editor NEW[ICLR'21] FedBN: Federated Learning on Non-IID Features via Local Batch Normalization
[ICLR'21] FedBN: Federated Learning on Non-IID Features via Local Batch Normalization
Under data_utils.py code,
" self.paths, self.text_labels = np.load('../data/office_caltech_10/{}_train.pkl'.format(site), allow_pickle=True)"
Can you provide the code to create .pkl file of the dataset used here?
Hi, I came across your work. It was an interesting idea, but it ran into some problems during the experiment.
When I clone your code, run the command below. The accuracy in the paper cannot be obtained.
# benchmark experiment
python fed_digits.py --mode fedbn --iters 300
This is my best experiment.
============ Train epoch 208 ============
MNIST | Train Loss: 0.0003 | Train Acc: 1.0000
SVHN | Train Loss: 0.0010 | Train Acc: 1.0000
USPS | Train Loss: 0.0004 | Train Acc: 1.0000
SynthDigits| Train Loss: 0.0007 | Train Acc: 1.0000
MNIST-M | Train Loss: 0.0007 | Train Acc: 1.0000
MNIST | Test Loss: 0.0985 | Test Acc: 0.9689
SVHN | Test Loss: 0.9124 | Test Acc: 0.7232
USPS | Test Loss: 0.1033 | Test Acc: 0.9683
SynthDigits| Test Loss: 0.5342 | Test Acc: 0.8359
MNIST-M | Test Loss: 0.6474 | Test Acc: 0.7904
Did I miss any details?
My platform
Ubuntu 20.04
GeForce RTX 2080 Ti
Intel(R) Xeon(R) Silver 4214 CPU @ 2.20GHz
Thanks for taking the time to help!
I have tried various methods to test the code of fedbn (use the trained model, input a picture for prediction), fedbn can not give the correct result, but fedavg is normal
Dear Authors,
How long does this model need to train?
we find that the train sets are kept the same size on all domains as the 'min_data_len' and we think there is a mistake:
FedBN/singleset/single_domainnet.py
Line 120 in 1caf83c
Hi, I have read your paper and code and found it to be an interesting work!
I have a question here, that is, why not test the performance of server model in the experiments?
If only the performance of client model is needed, it seems that it would be better train the model locally?
When I run ./utils/data_preprocess.py python An error has occurred:requests.exceptions.SSLError: HTTPSConnectionPool(host='www.baidu.com', port=443): Max retries exceeded with url: / (Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:997)')))
How to solve this error?
Would you please provide a preprocessed ABIDE I data set?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.