something wrong in FedML/fedml_api/standalone/hierarchical_fl group.py and trainer.py

Open cccrick opened this issue 4 years ago • 1 comments

In FedML/fedml_api/standalone/hierarchical_fl/group.py line 4， from fedml_api.standalone.fedavg.fedavg_trainer import FedAvgTrainer don't exist. I have tried to modify to my_model_trainer.py in fedml_api.standalone.fedavg, the params is error. And modified to FedAVGTrainer.py in fedml_api/distributed/fedavg, same error.

The command in the CI-script-fedavg.sh of hierarchical-fl is wrong. The command is follow

cd ./fedml_experiments/standalone/hierarchical_fl
sh run_standalone_pytorch.sh 0 1000 1000 -1 mnist ./../../../data/mnist lr hetero 0.03 sgd random 2 5 2 1

The same problem exist in the trainerl.py.

Oct 25 '21 02:10 cccrick

I think this is because of our recent code refactoring in other algorithms. We are doing FedML 2.0, will solve this kind of dependency issue systematically.

Oct 25 '21 04:10 chaoyanghe

Check our latest examples at: https://github.com/FedML-AI/FedML/tree/master/python/examples

We've upgraded our library a lot in recent version. Here is a brief introduction: https://medium.com/@FedML/fedml-ai-platform-releases-the-worlds-federated-learning-open-platform-on-public-cloud-with-an-8024e68a70b6

Aug 17 '22 01:08 chaoyanghe