chenxin061 pdarts issues

chenxin061 / pdarts

Codes for our paper "Progressive Differentiable Architecture Search:Bridging the Depth Gap between Search and Evaluation"

Other

359 stars 83 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

How to caculate GPU-days(cost)?

#41 liqier opened 3 years ago
0
The number of filters in each cell

#40 DaweiXi360 closed 3 years ago
0
Reproducing the Results

#39 kunjing96 opened 3 years ago
0
Reproducing the Results

#38 kunjing96 closed 3 years ago
0
CIFAR-100 Performance and Hyperparameters

#37 jyfliu opened 3 years ago
0
What does 'operation-level Dropout' mean?

#36 dk-hong opened 3 years ago
0
Question regarding the paper

#35 yxchng opened 3 years ago
1
PDARTS on medical images classification

#34 AissamDjahnine closed 4 years ago
1
parameter stem_multiplier and module stem

#33 xjhcassy closed 3 years ago
6
lr being passed to train(), but not used

#32 rtarquini closed 4 years ago
1
semantic segmentation

#31 NdaAzr closed 4 years ago
1
Is this a normal case?

#30 mrk1992 closed 4 years ago
0
Question about Dropout Rate in Paper

#29 volgachen closed 4 years ago
4
confusion about depths of network in different stages

#28 NdaAzr closed 4 years ago
1
How many layers we need in the evaluation stage?

#27 NdaAzr closed 4 years ago
11
In the original implementation of DARTS, it is input_search, target_search = next(iter(valid_queue), which slows down the training when using PyTorch 0.4 and above.

#26 BangguWu closed 4 years ago
1
关于一阶优化

#25 hyqyoung closed 4 years ago
1
correct RUN command

#24 xiebangquan closed 4 years ago
1
Is your method only work on CUDA 10?

#23 kids0cn opened 4 years ago
2
How to use this algorithm to my datasets..

#22 suryakumara closed 4 years ago
1
which genotype is best？

#21 johsnows opened 4 years ago
5
use train_search , got genotype like DARTS

#20 ykk648 closed 4 years ago
2
How can resume the train_search.py using weights in log_dir if the code stops due to accidents like power cut ... ?

#19 JarveeLee closed 4 years ago
1
In arch search process , why use SGD for operation weight and Adam for arch_params ?

#18 JarveeLee closed 4 years ago
1
How can i import the found network architecture ?

#17 cristyioan2000 opened 4 years ago
3
RuntimeError: cuda runtime error (2)

#16 liu-ca closed 4 years ago
0
if i need train it first and than evalutate?

#15 johsnows closed 4 years ago
4
why the learning rate and batchsize in "train_imagenet.py" are different from what said in the paper? Which one can reproduce the result？And is the learning rate related to batchsize ?

#14 AIshenfeng closed 4 years ago
1
For other datasets, is it better to use pre-searched network from cifar, or search from itself?

#13 rtrobin opened 5 years ago
13
[BUG] test.py

#12 zihaozhang9 closed 5 years ago
1
How to test?

#11 zihaozhang9 closed 5 years ago
4
[Bug]CUDA error: out of memory

#10 zihaozhang9 closed 5 years ago
1
How many hours on ImageNet of training?

#9 betterhalfwzm opened 5 years ago
11
Problem while running test.py

#8 Catosine closed 5 years ago
4
Reproducing the Results, and Questions

#7 drcdr opened 5 years ago
22
Add more code comment？

#6 coolKeen opened 5 years ago
3
Multi-gpu support

#5 R00Kie-Liu opened 5 years ago
10
Missing References

#4 D-X-Y closed 5 years ago
2
Can i control the flops of the models searched?

#3 zyc4me opened 5 years ago
3
如果只有ＧＴＸ１０８０ｔｉ（１１Ｇ内存），如何调整参数，让pdarts也可以跑起来训练？

#2 simonggx closed 5 years ago
1
Update README.md

#1 198808xc closed 5 years ago
1