sseung0703 KD_methods_with_TF issues - Githubissues

sseung0703 / KD_methods_with_TF

Knowledge distillation methods implemented with Tensorflow (now there are 11 (+1) methods, and will be added more.)

MIT License

266 stars 61 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

The results of Soft Logits fluctuate quite a lot

#18 zhongshaoyy opened 5 years ago
3
probably mistaken implementation of RKD methond

#17 wisdom0530 opened 5 years ago
3
Why only update on train_op2 but not train_op?

#16 snownus closed 5 years ago
4
Choice of SVD gradient

#15 jonkoi closed 5 years ago
1
which is the paper for KD_EID?

#14 guanfuchen closed 5 years ago
1
Create LICENSE

#13 sseung0703 closed 5 years ago
0
Can you provide the plot script?

#12 guanfuchen closed 5 years ago
1
How many epochs did you set to train the teacher model?

#11 Xiaocong6 opened 5 years ago
2
I found a major issue on network and distillation module!

#10 sseung0703 closed 5 years ago
1
Multi label classification

#9 adavoudi closed 5 years ago
1
hi, i think there were two more different methods for KD

#8 yyht closed 5 years ago
2
What about TF 2.0?

#7 Oktai15 closed 5 years ago
1
What dataset is used for the experiment?

#6 Xiaocong6 closed 5 years ago
2
Implement of ResNet

#5 cupwater closed 5 years ago
11
TypeError: ResNet() missing 1 required positional argument: 'label'

#4 CBHealth closed 5 years ago
9
Update train_w_distill.py

#3 CBHealth closed 5 years ago
1
AB algorithm fixed

#2 bhheo closed 5 years ago
0
Can you consider adding the method of attention transfer and neuron-selectivity-transfer?

#1 Xiaocong6 closed 5 years ago
4