Yonghongwei Gradient-Centralization issues - Githubissues

Yonghongwei / Gradient-Centralization

A New Optimization Technique for Deep Neural Networks

533 stars 80 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

请问该方法在以transformer为主结构的目标检测任务中有效吗

#15 wokeyide1999 opened 2 years ago
0
licence

#14 kornellewy opened 2 years ago
0
Question regarding GC for convolutions

#13 pavlin-policar closed 3 years ago
1
可以直接使用在3D卷积网络里吗？

#12 huamgmin closed 3 years ago
1
没有提升issue

#11 ZangHuanyu opened 4 years ago
7
CG方法只能针对conv层和FC层做梯度中心化吗？

#10 happy2019-k closed 4 years ago
2
incorrect keywords: memory_format

#9 jin-s13 opened 4 years ago
2
关于语义分割的问题

#8 GewelsJI opened 4 years ago
3
训练非常慢

#7 you-old opened 4 years ago
3
Consider RAdam as the successor of Adam

#6 LifeIsStrange closed 4 years ago
2
Should i use pytorch gradient clippping with gradient centralization?

#5 hadaev8 closed 4 years ago
2
Is it ok to use GC with weight decay in style of AdamW?

#4 hadaev8 closed 4 years ago
2
Does this work for Conv1d?

#3 Oktai15 closed 4 years ago
1
function grad.mean() Error

#2 lucaskyle closed 4 years ago
1
Update main.py

#1 Zhaoyi-Yan closed 4 years ago
0