Image summary is lost after ds_iterator.get_next()

[ghost]

Hi, thanks for the wonderful code!

I found the summaries added during image preprocessing are lost after calling ds_iterator.get_next().

I identify this issues by adding tmp = tf.get_collection(key=tf.GraphKeys.SUMMARIES) right after ds_iterator.get_next(). The returned value of tmp is an enmpty list. However, when adding the same line at the end of def preprocess(self, raw_image), the returned value of tmp is the summary proto for the processed images, which is normal.

I, therefore, guess that the summaries added before calling Dataset iterator will not be retained. Is this true? If so, do we have a way to circumvent this issue?

Many thanks.

[tfboyd]

I want to ack that I saw this issue. I am not sure anyone will look at it. We do not use the summaries very often in the benchmark. I will try to get someone to take a quick look. While not a great answer or super useful, I prefer to be transparent rather than just ignore the feedback and question.

[ghost]

Thanks anyways. Your effort is truly appreciated~ ^_^

[tfboyd]

It may get fixed. I hate being a downer. So much to do and not enough time.

On Oct 24, 2017 5:36 PM, "YusongLeng" [email protected] wrote:

Thanks anyways. Your effort is truly appreciated~ ^_^

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/tensorflow/benchmarks/issues/69#issuecomment-339177496, or mute the thread https://github.com/notifications/unsubscribe-auth/AWZesjOyF1kt00wHwfJV2oLIYTOJbFJBks5svoKRgaJpZM4QAMfQ .