search

melodyguan / enas

TensorFlow Code for paper "Efficient Neural Architecture Search via Parameter Sharing"
https://arxiv.org/abs/1802.03268
Apache License 2.0
1.58k stars 390 forks source link

Questions about macro search #62

Open HanDarkholme opened 6 years ago

HanDarkholme commented 6 years ago

After RNN training, in your code, you do not choose the one with the best acc to retrain the corresponding CNN, I wonder how to explain this.