Tensorflow version problems

[DreamingButterfly]

Hello, thanks for your nice work on graph classification.
Recently, when I use your code to run experiments, I use pip install -r requirement.txt, but pip tells me that tensorflow1.15 requires tensorboard>1.15 <1.16, but tensorflow-gpu2.3 needs tensorboard>2.3. So I am eager to know how to build a env with correct tf version to run your code, thanks.

[Yang Beining]

It is common problem, especilly in tensorflow (abandon reason ++). I freeze the other env list, and test it, it looks like working good. Hope it can help you.
requirements.txt

[DreamingButterfly]

Thanks. I solved it with using tf2.2, cause that tf1.14 needs cudnn7.4 & cuda 10.0, which is uncompatible for my machine. And this is my requirements.txt, who use tf2.x can refer to it.
And for the problem of migrating tf1.x code to tf2.x, please read tf migrate. As for tf.contrib, delete line129, 130 in train.py, and change line 135, 136 in train.py from
self.optimizer = tf.compat.v1.train.MomentumOptimizer(self.lr, self.mom).minimize(self.loss + self.l2)
to
self.optimizer = tf.compat.v1.train.MomentumOptimizer(self.lr, self.mom).minimize(self.loss +tf.compat.v1.losses.get_regularization_losses()).
So far, tf2.x users can run the code succefully.