Hello,
I’ve been struggling on an error stating “Wrong value in outd”.
Note 1: All tests passed for “DecoderLayer_test(DecoderLayer, create_look_ahead_mask)” Note 2: I used training=training in the dropout layer for both the DEcoderLayer and the Decoder
Any hint to share? Thank you
** Update: I was wrongfully using tf.math.square instead of tf.math.sqrt…
Everything works well now.
My take away is working on his own without people interactions, especially to have fresh pairs of eyes to spot dumb errors, is quite challenging…
Hi @bblanc
So glad that you find out.
Keep learning!
