How to load pre-trained weights before starting the run?

[ajulyav]

Hello,

I am not sure I got the idea how can I start with the pre-trained model. I have .pt file which I sent to all clients.
As well, I have added "source_ckpt_file_full_name": "/media/Experiment/NVFlare/workspace/init_model/init_global_model.pt"
to config_fed_server.json.

I am using the template of Prostate project. Should I add the lines to supervised_learner.py?
I appreciate your help in advance or any tutorial in future.

Thank you!

[holgerroth]

The server will be sending the initial weights for each FL round. If your persistor in config_fed_server.json loads the checkpoint, it means those weights will be sent to each client in the first round. You don't need to modify supervised_learner.py. CC @ZiyueXu77