enas

enas copied to clipboard

how is the effect on larger images (e.g. 300*300)

1

Hi, I have run your code on the dataset of flower, whose image size is much larger than Cifar. The result is not very good, only about 80% accuracy. I'm...

meiniangpp

Errors in attempt at reproducing micro search

3

I tried running micro search on TF 1.7 and it made quite a bit of progress, up to 150 epochs, but then it failed out as follows: ``` [1 2...

ahundt

Hi, Thanks for the great work and sharing the implementation to the community! I recently found that this part of codes looks kind of weird. During micro-search for CIFAR-10, you...

hubert0527

RNN results not reproducible

2

Hello, Thanks for open sourcing the code. After your commit: https://github.com/melodyguan/enas/commit/2734eb2657847f090e1bc5c51c2b9cbf0be51887 I get 63.26 in ppl and not the 55.6 stated in the paper. However before this commit I get...

Wronskia

Allows optional usage: `$ ptb_final.sh "0 0 0 1 1 2 1 2 0 2 0 5 1 1 2 6 1 8 1 8 1 8 1"`

philtomson

Is it possible to apply 1D data to this method? I tried to just change the size of the filters of convolution and also the size of pooling, but ending...

kentropy

Error message: slice index 0 of dimension 0 out of bounds.

1

Hello, Thank you for your interesting paper. While executing the code, I encountered the error below: File: general_child.py Line: 362 - res_layers.append(tf.cond(tf.equal(skip[i], 1), lambda: prev_layers[i], lambda: tf.zeros_like(prev_layers[i]))). ValueError: slice index...

heraldiclily

Runtimes and results

1

In section 3.1 Results in the paper it says "Running on a single Nvidia GTX 1080Ti GPU, ENAS finds a recurrent cell in about 10 hours." I ran th e...

philtomson

Reproducibility of the results from the paper (RNN)

11

Hi @melodyguan, thanks for the great paper! But I still can't reproduce results from your paper in finding the RNN cell on PTB dataset. After approx. 24 hours of training...

nikitos9000

use which script can reproduce the result shown in the paper?

1

I can see the many parameters in the script? while some parameter are different from the lines in the paper I want to reproduce the result shown in the paper?...

yogurfrul