forever1078

RuntimeError: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 3.94 GiB total capacity; 2.65 GiB already allocated; 2.25 MiB free; 4.11 MiB cached)

5

``` Epoch: 0 Traceback (most recent call last): File "main.py", line 142, in train(epoch) File "main.py", line 93, in train outputs = net(inputs) File "/home/.local/lib/python3.5/site-packages/torch/nn/modules/module.py", line 493, in __call__ result...

In FederatedAveragingStrategy class processSingleUpdate function,why the normalisedUpdate equals to update div normaliser? normaliser=samples/totalsamples normalisedUpdate=update/normaliser=update*totalsamples/samples? I can't understand here.Are you using the FedAvg algorithm?