Flattens the filter to a 2-D matrix with shape [filter_height * filter_width * in_channels, output_channels] . · Extracts image patches from the input tensor to ...
確定! 回上一頁