torch.nn.init. calculate_gain (nonlinearity, param=None)[source]. Return the recommended gain value for the given nonlinearity function.
確定! 回上一頁