how to change the network module in batch_eth_mnist

[MayWeHappy]

I encountered an error when running the network.run function after replacing the network in batch_eth_mnist with DiehlAndCook2015v2. The error occurred in the learning.py file at line 399, specifically in the _connection_update method: self.connection.w -= self.reduction(torch.bmm(source_s, target_x), dim=0).

The error message states that the output shape, [784, 100], does not match the broadcast shape, [32, 784, 100]. I tried using batch processing in other network training routines, such as SOM, but encountered the same issue.

Could you please advise on how to resolve this issue? Thank you very much.

[Hananel-Hazan]

The network DiehlAndCook2015v2 is not designed to work with batch size. To adapt the network you need to modify the shape of the network input to match the batch size