mynameisissac

As described in the paper, the 'running_mean_bmic' and 'running_std_bmic' are used to compute the client-sharing statistics variances 'running_var_mean_bmic' and 'running_var_std_bmic'. This is done as expected in the code. But in...

https://github.com/JeremyCJM/CCST/blob/87c3e448cfa28c366b58d17f8dffc5d3d83574e5/federated/fed_run.py#L112-L122 Your meta learning implementation is wrong, `load_state_dict()` does not preserve the computation graph so `loss_outer` will never propagate gradient back to the original model parameters. be careful :)