Instead of explicitly adding a new dimension, in tensorflow 2.7.0 you could just use keepdims=True as an argument of the ...
確定! 回上一頁