Hieu Pham

**About your derivations.** I do not see anything wrong with @dgedon's derivation. > `f(x) = f(a) + f'(a)(x-a)`. > So there is f'(a) instead of f'(x). Then with the same...

Hi @nsmetanin, Thank you for your interest in our work. Given the description of your experiments, we suspect that you ran the script `ptb_search.sh`, and not the `ptb_final.sh`. The script...

@nsmetanin Thanks! Please let us know how it goes 😃

Hi Hanxiao @quark0, That's definitely a bug. Thank you for spotting it. We have pushed [a commit](https://github.com/melodyguan/enas/blob/master/src/ptb/ptb_enas_child.py#L480) that fixes it. We have tried rerunning the code and the output looks...

**Update on results:** @quark0 We finished rerunning the script with the fix and indeed got the test perplexity of `56.6`.

Yes that's what we got too. We think the reason is that the validation perplexity we computed using a `batch_size` of `35`. If we use `batch_size = 1` for validation,...

Hi, Thanks for your interest. Commit [2734eb2](https://github.com/melodyguan/enas/commit/2734eb2657847f090e1bc5c51c2b9cbf0be51887) fixed a bug in the evaluation process. After the fix, we had to further tune the model's hyper-parameters to reach a good performance....

``` # Exp. 1 ./scripts/ptb_search.sh # should give you a bunch of architectures ./scripts/ptb_final.sh # should give you around 55.8 test perplexity on PTB # Exp. 2 ./scripts/cifar10_macro_search.sh # should...

Thank you for dedicating the time and resource to run and verify our code. We also looked at the architectures that were sampled in the time steps before and took...

Thanks for noting this. We just re-ran the script and didn't see the issue. We use TensorFlow 1.4.1 though. We'll update to 1.7 soon and let you know if we...