Bug after recent Tensorflow update

[adai183]

cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits, one_hot_y)
need to be:
cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=one_hot_y)

If not we get:

Raise en error if labels do not sum to 1 exception

This came up in the discussion forum.

[mwhayford]

Thanks for this comment! I had this same issue to when loading for the first time.

[nandanlahurikar]

cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=LeNet(x), labels=one_hot_y)

[zippau]

agree... cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=one_hot_y)

[CreatCodeBuild]

I am a classroom mentor. New students start to get this bug because they use 1.0.

I think Udacity code should use 1.0 standard as well.

[CreatCodeBuild]

@shaox058, the best way it to modify your code. It's trivial to do so. You just need keyword arguments.

…

On Sun, Mar 26, 2017 at 12:57 AM shaox058 ***@***.***> wrote: Hi @CreatCodeBuild <https://github.com/CreatCodeBuild> @adai183 <https://github.com/adai183> @mwhayford <https://github.com/mwhayford> I have the same issue here. Should I downgrade my tensorflow? If yes, which version is the best fit. Thanks, Harry — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#16 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AIga8sKFUyrWaVvjJu1fLunllb6jSfuPks5rpfBEgaJpZM4MKJR1> .

[mvirgo]

Since we are still using v0.12.1, closing this - will be updated when we move to the newer versions.