Comments (12)
midasklr
commented on August 22, 2024
换主干网络改的地方多吗？比如mobilenetv2
改起来很简单，可以参考yolo.py里面解析模型那部分。
我换过mobilenetv3 small，在coco上掉点比yolov5s多很多。
from yolov5prune.
sjfei
commented on August 22, 2024
换主干网络改的地方多吗？比如mobilenetv2
改起来很简单，可以参考yolo.py里面解析模型那部分。
我换过mobilenetv3 small，在coco上掉点比yolov5s多很多。
我自己的数据集，感觉还可以，就是剪枝的时候报错，想问下剪枝的话需要怎么改一下
from yolov5prune.
midasklr
commented on August 22, 2024
可以用轻量的骨干然后仅仅对head裁剪。
from yolov5prune.
sjfei
commented on August 22, 2024
可以用轻量的骨干然后仅仅对head裁剪。
我也这样想的，但是不知道怎么改一下，让它只剪枝head
from yolov5prune.
midasklr
commented on August 22, 2024
可以用轻量的骨干然后仅仅对head裁剪。
我也这样想的，但是不知道怎么改一下，让它只剪枝head
见:
Line 360 in b314ca0
这里model_list字典就是需要剪枝的层,你可以看看key然后去除backbone部分就行
from yolov5prune.
sjfei
commented on August 22, 2024
可以用轻量的骨干然后仅仅对head裁剪。
我也这样想的，但是不知道怎么改一下，让它只剪枝head
见:
Line 360 in b314ca0
这里model_list字典就是需要剪枝的层,你可以看看key然后去除backbone部分就行
大佬，我把backbone部分都删除了，但是还是有错误，并且程序还把bockbone剪枝了，不是很懂
报错
from yolov5prune.
midasklr
commented on August 22, 2024
部分层的remaining channel等于0了 ，设置更小的剪枝比例，或者限制每一层至少保留一个通道。
from yolov5prune.
sjfei
commented on August 22, 2024
部分层的remaining channel等于0了 ，设置更小的剪枝比例，或者限制每一层至少保留一个通道。
大佬，设置小了还是会还报错，是不是让它只剪枝head的代码修改有问题，报错和上面一样
from yolov5prune.
midasklr
commented on August 22, 2024
等这两训练完了 我提交一下 yolov5s-mobilenetv3_small_0.75实验结果和代码
from yolov5prune.
sjfei
commented on August 22, 2024
等这两训练完了 我提交一下 yolov5s-mobilenetv3_small_0.75实验结果和代码
好的，感谢大佬了
from yolov5prune.
sjfei
commented on August 22, 2024
等这两训练完了 我提交一下 yolov5s-mobilenetv3_small_0.75实验结果和代码
大佬，想问一下什么时候更新代码
from yolov5prune.
youbearlai2015
commented on August 22, 2024
同问
from yolov5prune.
Related Issues (20)
- 我想请教下稀疏训练的稀疏效果一直不好是怎么回事，我甚至把sr开到了10 HOT 2
- prune.py脚本运行报错 too many indices for tensor demension 3 HOT 1
- 计算量 HOT 1
- RuntimeError: mat1 and mat2 shapes cannot be multiplied HOT 1
- 训练自己的算法
- 使用train_sparsity.py出现的错误- HOT 1
- 已经充分稀疏的情况下，剪枝率只能设置到0.1
- 关于train.py训练后tensorboard中没有bn_weights/hist直方图 HOT 1
- v7.0版本能用吗？ HOT 4
- 运行预测脚本的时候报错了 显示yolo脚本没有DetectionModel这个属性
- 更换主干网络再经行稀疏化训练后，值变成了nan
- tensorboard中的histogram图表中的数据如何导出呢
- IndexError: too many indices for tensor of dimension 3
- 稀疏因子最高只能设为0.1,再高一点剪枝最大比例反而会降低是怎么回事
- 剪枝出现Scanning 'E:\task\experiment\dataset\BITVehicle\BITVehicle\val.cache' images and labels... 0 found, 985 missing, 0 empty, 0 corrupted HOT 1
- 博主麻烦看看这个问题 进行prune.py 出现KeyError: 'model.0.bn'问题
- Can't get attribute 'ModelPruned' on <module 'models.yolo'
- gamma, bias distribution
- 剪枝完之后在GPU上的推理速度反倒比剪枝前的模型更慢
- 估计是裁剪完的网络通道数不是2^n，导致再GPU上并行加速效果下降，推理时间就增加了。CPU上应该会有速度提升。 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 yolov5prune.