Eric-mingjie network-slimming issues

Eric-mingjie / network-slimming

Network Slimming (Pytorch) (ICCV 2017)

MIT License

909 stars 214 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Regarding the calculation of FLOPs after model compression

#93 mjw123bs opened 8 months ago
1
对HRNet剪枝，第一个bn的grad是None的问题

#92 zhangdandan-git opened 10 months ago
0
ResNet和DenseNet这种每一层的输入会作为后续多个层的输入，且其BN层是在卷积层之前，在这种情况下，稀疏化是在层的输入末端得到的，一个层选择性接受所有通道的子集去做下一步的卷积运算。

#91 NingBoB opened 10 months ago
0
题外话：模型压缩如何入门？对于自己的网络架构该如何着手去写剪枝代码？

#90 XXMxxm220 opened 1 year ago
0
About other visions

#89 panhu opened 1 year ago
0
RuntimeError: CUDA error: device-side assert triggered

#88 csn223355 opened 1 year ago
1
问题咨询：剪枝后通道数为0

#87 zhangtao1234567 opened 1 year ago
7
m.weight.grad.data.add_的问题

#86 alexxiecan opened 1 year ago
0
关于L1 regular

#85 alexxiecan opened 1 year ago
1
经稀疏训练剪枝后模型变小，但是refine微调后模型又变大了

#84 ttttttt12 opened 2 years ago
0
mnn加载剪枝模型错误

#83 DLTensor opened 2 years ago
1
TypeError: item() takes no arguments (1 given)

#82 Rhett-Lin opened 2 years ago
1
原论文中的memory是指

#81 PQ-girl opened 2 years ago
6
If the remaining channel for a layer is zero, it reports zero division error

#80 Neil-untitled opened 2 years ago
2
RuntimeError: Given groups=1, weight of size [15, 14, 1, 1], expected input[64, 16, 32, 32] to have 14 channels, but got 16 channels instead

#79 lovegit2021 opened 2 years ago
1
剪枝后保存的权重文件和newmodel加载不上

#78 lovegit2021 opened 2 years ago
3
channel_selection layer intraining process

#77 wendychina opened 2 years ago
2
Sparse Confusion

#76 EddieEduardo closed 2 years ago
1
Minor Bugs caused by old version

#75 planemanner opened 3 years ago
0
预测单张照片时，准确率只有0.0几

#74 PQ-girl opened 3 years ago
2
为什么我裁剪完成后，模型运行占用的显存比没裁剪的还要高？

#73 LW-CVer opened 3 years ago
1
为什么定义的resnet网络训练时gpu显存占用这么大啊？

#72 LW-CVer closed 3 years ago
0
resnet50

#71 creatonce opened 3 years ago
1
关于thersh

#70 L0-zhang closed 3 years ago
2
关于channel select的mask剪枝方式

#69 fire717 opened 3 years ago
2
请问为何微调之后保存的模型大小变大了

#68 wayne2tech opened 3 years ago
4
用自己的数据集训练

#67 PQ-girl opened 3 years ago
3
about resnet20 slimming

#66 Danee-wawawa opened 3 years ago
1
Results are not replicable even with VGG19. I guess details mentioned in Readme.md is no where a match.

#65 chowkamlee81 opened 3 years ago
1
关于剪枝percent的疑问

#64 nanabibi1204 opened 3 years ago
1
弱弱的问下，ResNet164这个模型哪里下的。。

#63 fsj0122 opened 3 years ago
12
the questions in resprune.py

#62 kingoo123 opened 3 years ago
5
thrresprune

#61 kingoo123 closed 3 years ago
0
关于FLOPs计算

#60 arthur0219 closed 3 years ago
2
卷积BN融合

#59 tangchen2 opened 3 years ago
2
关于文件保存

#58 arthur0219 closed 3 years ago
1
there is no `resume` is defined.

#57 erhsh opened 3 years ago
1
IndexError: index 0 is out of bounds for dimension 0 with size 0

#56 jiagnhaiyang opened 4 years ago
3
Testing on SVHN

#55 kbui1993 closed 3 years ago
1
A tough question about network-slimming used on SSD

#54 windyjunfeng opened 4 years ago
1
关于BN层bias的裁剪

#53 wxy1234567 opened 4 years ago
4
关于剪枝代码的疑虑

#52 Ixuanzhang opened 4 years ago
5
gamma参数相关问题

#51 zhongqiqianga opened 4 years ago
5
关于channel selection layer的问题请教

#50 YuQi9797 opened 4 years ago
4
关于论文中稀疏训练的损失函数的疑惑

#49 songyang86 opened 4 years ago
3
Resnet中bn和conv的位置关系问题？

#48 Curry-Christopher opened 4 years ago
1
为什么剪枝生成的tar模型压缩包打不开啊？无法解压，显示损坏

#47 pingzi5233 closed 4 years ago
2
What if the case that gamma is near zero, but beta is very large?

#46 yzlyty opened 4 years ago
1
Why not stop gradient for channel_selection layer's parameters?

#45 yzlyty opened 4 years ago
1
请教下100分类中稀疏度的值

#44 xuemo188 opened 4 years ago
0