# From https://www.tensorflow.org/versions/r2.9/api_docs/python/tf/GradientTape#gradient
x = tf.ragged.constant([[1.0, 2.0], [3.0]])
with tf.GradientTape() as g:
g.watch(x)
y = x * x # HERE
g.gradient(y, x)
Consider the multiplication expression at "HERE" above. In this context, x is a tensor, and multiplying two tensors yields a tensor. But, I cannot see a way to express that in tensorflow.xml.
Consider the following code:
Consider the multiplication expression at "HERE" above. In this context,
x
is a tensor, and multiplying two tensors yields a tensor. But, I cannot see a way to express that intensorflow.xml
.Regression
The function
tf.multiply()
seems to work fine.