You could get the running mean and var (exponentially weighted average from last batches) with bn.running_mean and bn.running_var , if that ...
確定! 回上一頁