NK

Improve LeakGAN by Changing Policy Gradient Structure

5

Hey @CR-Gjx Thanks for providing this open source code. Very helpful to study and I love the idea of hierarchical reinforcement learning. In the recent AlphaGo Zero paper and [Thinking...

SRU Module Doesn't appear to Use Residual/skip connections

5

@taolei87 thanks for the repo again. Really good code. One thing that I was analyzing was that it doesn't seem that the sru class doesn't have skip connections. Shouldn't it...

Use this code on another repo?

1

Hey Oliver, Great job implementing conv lstm. Someone requested me to add this type of LSTM to my repo: https://github.com/LeavesBreathe/tensorflow_with_latest_papers Just wondering if we can copy and use your code...

Errors on Using .so Files

21

Hey Chiu, Unfortunately getting an error when I run both .so files: `tensorflow.python.framework.errors_impl.NotFoundError: layer_norm_fused_op.so: undefined symbol: _ZN10tensorflow8internal10LogMessage12MinVLogLevelEv ` I could try to dig in and modify the c++ code but...

Padding Affecting Batch Norm

12

Hey Tim, The batch norm LSTM paper published has pretty stellar results. I've followed the keras issue thread where you stated that gamma and beta are shared throughout all timesteps,...

Multiple Heads for In and Out Keys

2

Thanks for publishing your code for this paper! One portion of your paper talks about using multiple heads for solving the arithmetic challenge. For calculate the `next_cells` did you simply...

Applying Weight Normalization to Single Vector Settings

2

Hey Tim, Thanks for publishing this code and it has been very useful for implementation purposes. I wanted to ask you about applying to weight norm to single vector settings....

Backproping Through Argmax

7

Hey Liamb, really appreciate the work you're doing here. For a long time, I've wanted to apply Adversarial networks to NLP. The main problem is that in a seq2seq generator,...

Interpreting Critic Loss To Improve Convergence

8

Hey @igul222, thanks again for your help with comments earlier. I've been rereading both WGAN papers, and I understand that the critic loss is supposed to estimate the wasserstein distance....

Hey @igul222 , just wanted to add a small update. I've found that if you add the historical buffer as Apple detailed [here](https://machinelearning.apple.com/2017/07/07/GAN.html), you can increase performance slightly and convergence...