ENAS-pytorch

ENAS-pytorch copied to clipboard

Retraining from scratch yields worse results

5

Hello, As written in the paper (2.2 _Deriving Architectures_) I tried to re-retrain from scratch the best derived model, but it surprisingly gives worse result when I retrain it from...

nkcr

Cannot use test and reproduce the result?

7

@dukebw Hi, thanks for your code. I download it and run it but meet with 3 main problems: 1. It seems that it cannot reproduce the result in the paper?...

renqianluo

I notice that in Trainer.train() self.save_model() is called at certain times to save all of the shared weights in the "super graph" (my terminology), but I don't see that the...

philtomson

SyntaxError: invalid syntax

3

hello, guys. I have implemented **ENAS-pytorch** on my computer. When I run this command: `python main.py --network_type rnn --dataset ptb --controller_optim adam --controller_lr 0.00035 --shared_optim sgd --shared_lr 20.0 --entropy_coeff 0.0001`...

zhjmcjk

Hey, Thanks a lot for implementing ENAS in Pytorch! I am able to run ENAS with Penn Treebank dataset. I am trying to understand the output during training. At beginning,...

yuezha01

Can you please list the versions of all the dependencies of this repository?

roeemz

Thanks for your implementation! May I ask for the command to retrain the searched model like ./scripts/ptb_final.sh in original repo? Or is this feature not yet implemented?

Chen-Hailin

CNN

12

carpedm20

Hidden State of RNN (not averaged ?)

2

Hello guys , Thank you very much for this work 👍 ! I have a quick question about the code : Why in the cell method you output `output, h[self.args.num_blocks...

kamilbs

When I run this code by : python main.py --network_type cnn --dataset cifar --controller_optim momentum --controller_lr_cosine=True \ --controller_lr_max 0.05 --controller_lr_min 0.0001 --entropy_coeff 0.1 I get some errors. Would you please...

axiniu