Lescurel
2020-11-30 16:57:07
You can use tf.strings.to_number :
>>> numbers_as_string = tf.constant(list("123456789"))
>>> numbers_as_string
<tf.Tensor: shape=(9,), dtype=string, numpy=array([b'1', b'2', b'3', b'4', b'5', b'6', b'7', b'8', b'9'], dtype=object)>
>>> tf.strings.to_number(numbers_as_string)
<tf.Tensor: shape=(9,), dtype=float32, numpy=array([1., 2., 3., 4., 5., 6., 7., 8., 9.], dtype=float32)>
For the merging, use tf.concat :
>>> a = tf.random.normal((2,100))
>>> b = tf.random.normal((2,50))
>>> a.shape
TensorShape([2, 100])
>>> b.shape
TensorShape([2, 50])
>>> tf.concat([a,b], axis=1)
<tf.Tensor 'concat:0' shape=(2, 150) dtype=float32>
热门帖子
热门github
6