Hello,
Does Matlab R2019a support Automatic Differentiation / Algorithmic Differentiation like TensorFlow's tf.gradients() in its Deep Learning Toolbox (or otherwise)?
I.e., a generic operator which accepts a (deep) network and returns another network which computes the gradients of the first one (with respect to inputs)?
This would be useful in implementing 2nd order networks.
Thx
0 Comments
Sign in to comment.