Comments (8)
不好意思标题打错了,是compress_rate
from hrank.
您好!我有两个问题:
- compress_rate是指每层的保留率吗,还是指的是每层要删掉的比例呢?
- 是在开始剪枝前仅计算一次各卷积层的平均秩,还是每层剪枝完后计算下一个卷积层的平均秩呢(比如第1层剪完了,然后我才计算第2层的平均秩;然后再剪第2层,再计算第3层的平均秩......)?
- 删掉比例
- 剪枝前仅计算一次各卷积层的平均秩
from hrank.
感谢!再次叨扰您,我又有了新的问题:
-
vgg16的压缩率和flops以及参数量和readme种给的不对应?(vgg16的compress_rate [0.95]+[0.5]*6+[0.9]*4+[0.8]*2,readme给的是Flops=105.61M和Params=2.64M,但是好像不对应,和我算出来的不对应)
-
是每剪一层就微调一次,还是将所有层剪完之后只微调一次?如果是前者,那么在剪枝前仅计算一次各卷积层的平均秩会不会有偏差呢?
from hrank.
- 没有吧。计算代码我们都提供的好好的,自己详细阅读我们的说明
- 一层一微调。 都剪完再调参考 https://github.com/lmbxmu/HRankPlus
from hrank.
老师,您好!我想请问下,HRank在对不同的网络模型进行剪枝时,每层的剪枝比例是如何确定的?是数学方法的计算结果,还是通过参数调整试错得出的?
from hrank.
感谢!再次叨扰您,我又有了新的问题:
- vgg16的压缩率和flops以及参数量和readme种给的不对应?(vgg16的compress_rate [0.95]+[0.5]*6+[0.9]*4+[0.8]*2,readme给的是Flops=105.61M和Params=2.64M,但是好像不对应,和我算出来的不对应)
- 是每剪一层就微调一次,还是将所有层剪完之后只微调一次?如果是前者,那么在剪枝前仅计算一次各卷积层的平均秩会不会有偏差呢?
针对你的第一个问题,我算的也不是2.64M,是2.92M,还记得吗你算得是多少
from hrank.
感谢!再次叨扰您,我又有了新的问题:
- vgg16的压缩率和flops以及参数量和readme种给的不对应?(vgg16的compress_rate [0.95]+[0.5]*6+[0.9]*4+[0.8]*2,readme给的是Flops=105.61M和Params=2.64M,但是好像不对应,和我算出来的不对应)
- 是每剪一层就微调一次,还是将所有层剪完之后只微调一次?如果是前者,那么在剪枝前仅计算一次各卷积层的平均秩会不会有偏差呢?
针对你的第一个问题,我算的也不是2.64M,是2.92M,还记得吗你算得是多少
啊,params我手算的是0.702755 M;然后我用pytorch扩展的“thop”库的profile函数算的是0.70M(精度不同)
from hrank.
- 这个repo里面的确vgg16_bn算的和readme不一致. 原因是代码: https://github.com/lmbxmu/HRankPlus/blob/master/cal_flops_params.py#L36 我理解是少了最后一层的剪枝和fc计算量增加导致计算不一致.
from hrank.
Related Issues (20)
- 关于rank的疑问 HOT 2
- Download ResNet-50 on ImageNet HOT 2
- 代码疑问 HOT 2
- The meaning of the "rank" of feature maps. HOT 3
- 关于剪枝后模型的问题 HOT 1
- 关于Vgg16中compress_rate的问题 HOT 5
- 预训练模型下载链接
- Pruning other algorithmic models
- 论文中的公式问题 HOT 1
- rank_generation.py 中关于 参数 --gpu的疑问?
- 模型问题
- 学习率设置
- Size of the feature maps at later layers
- Automatically resume training from the highest test acc epoch may cause data leak.
- 有关计算flops和params的问题
- How to determine the per-layer filter pruning rate for a given model HOT 1
- Question about code
- Question about code? HOT 18
- 关于论文图2的问题 HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from hrank.