Comments (5)
containing BN and LeakyReLU?
from yolov3-model-pruning.
pruned_model是裁剪了的模型,filters还是跟原始模型一样,compact_model 修改了filters,需要给每一层的权重用init_weights_from_loose_model初始化一下
from yolov3-model-pruning.
请问pruned_model里面包括对BN的bias的裁剪吗
from yolov3-model-pruning.
@violet17 ,CBL_idx 是卷积层+BN+LeakyReLU的简写而已,C是指卷积
from yolov3-model-pruning.
@Lam1360 哦哦 谢谢哈,还有一个问题remaining channel 会不会只剩下一两个channel啊?我现在剪枝完之后就只剩下一两个channel。
layer index: 0 total channel: 32 remaining channel: 19
layer index: 2 total channel: 32 remaining channel: 10
layer index: 6 total channel: 64 remaining channel: 2
layer index: 9 total channel: 64 remaining channel: 24
layer index: 13 total channel: 128 remaining channel: 20
layer index: 16 total channel: 128 remaining channel: 36
layer index: 19 total channel: 128 remaining channel: 27
layer index: 22 total channel: 128 remaining channel: 24
layer index: 25 total channel: 128 remaining channel: 12
layer index: 28 total channel: 128 remaining channel: 11
layer index: 31 total channel: 128 remaining channel: 9
layer index: 34 total channel: 128 remaining channel: 13
layer index: 38 total channel: 256 remaining channel: 61
layer index: 41 total channel: 256 remaining channel: 41
layer index: 44 total channel: 256 remaining channel: 7
layer index: 47 total channel: 256 remaining channel: 10
layer index: 50 total channel: 256 remaining channel: 4
layer index: 53 total channel: 256 remaining channel: 2
layer index: 56 total channel: 256 remaining channel: 18
layer index: 59 total channel: 256 remaining channel: 5
layer index: 63 total channel: 512 remaining channel: 86
layer index: 66 total channel: 512 remaining channel: 7
layer index: 69 total channel: 512 remaining channel: 1
layer index: 72 total channel: 512 remaining channel: 1
layer index: 75 total channel: 512 remaining channel: 94
layer index: 76 total channel: 1024 remaining channel: 190
layer index: 77 total channel: 512 remaining channel: 83
layer index: 78 total channel: 1024 remaining channel: 186
layer index: 79 total channel: 512 remaining channel: 96
layer index: 80 total channel: 1024 remaining channel: 211
layer index: 87 total channel: 256 remaining channel: 59
layer index: 88 total channel: 512 remaining channel: 94
layer index: 89 total channel: 256 remaining channel: 37
layer index: 90 total channel: 512 remaining channel: 87
layer index: 91 total channel: 256 remaining channel: 52
layer index: 92 total channel: 512 remaining channel: 114
layer index: 99 total channel: 128 remaining channel: 29
layer index: 100 total channel: 256 remaining channel: 62
layer index: 101 total channel: 128 remaining channel: 23
layer index: 102 total channel: 256 remaining channel: 43
layer index: 103 total channel: 128 remaining channel: 21
layer index: 104 total channel: 256 remaining channel: 77
from yolov3-model-pruning.
Related Issues (20)
- 精度过低
- 能用于别的模型么?比如bisenet
- 推理速度 自测是22ms左右,不是7ms
- 请问如果训练多个类别,该如何配置?看标签文件没有写类别,是只能训练一类吗?
- 剪枝后FPS速度问题 HOT 2
- train.py训练完为啥没有选出一个最优的yolov3_ckpt.pth额,只是生成了一系列不同epoch的pth
- 请问这里用于稀疏化训练的预训练权重可以是之前训练好的模型吗? HOT 1
- "Channels would be all pruned!"
- 求FLOPS代码位置 HOT 1
- 减枝完,检测不到结果
- prune_model_keep_size 只是将model中冗余的通道通过mask置零,并没有真正的剪掉这些通道? HOT 1
- Cooperation Proposal: Lam1360/YOLOv3-model-pruning & PaddlePaddle
- 模型裁剪以后重新训练可以生成新的pth, 但是同时生成.weights文件报错。
- 这个模型适用于其他crnn结构嘛?
- 训练时候cls_acc一直是100% HOT 1
- 为什么上一层丢掉的权重,下一层的BN层runing_meaning要减去offset
- 为什么跑的这个数据集的结果只有map0.68,作者写的0.75 HOT 1
- 为什么会出现No such file or directory: 'data\\images\\train\\Buffy_1.jpg'
- 为什么剪枝后的 sample_metrics 为空 HOT 2
- 请问大家检测速度的代码在哪里啊?没有看到
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 yolov3-model-pruning.