Comments (8)
@QingL0218
The concrete values of loss do not matter. In our implemented version, the computation of metric is combined with loss layer. In our view, if the values can reflect the loss trend, the values are valid for metric.
from lffd-a-light-and-fast-face-detector-for-edge-devices.
@YonghaoHe 我在训练的时候,会出现loss为nan的情况,请问这是咋回事,我是用自己的数据集去训练的
from lffd-a-light-and-fast-face-detector-for-edge-devices.
@Dreamgang 我们在训练的过程中几乎没有出现过nan,印象里就出现过几次nan.
给你几个建议:
1,把学习率调小,比如从0.1->0.05,还有适当增加batchsize(32是我们常用的)
2,查看数据集里面是否有严重噪声数据,一些严重噪声数据会瞬间破坏参数的学习
from lffd-a-light-and-fast-face-detector-for-edge-devices.
@YonghaoHe 我在用V2训练时,batch_size设置32,学习率设0.1-0.05间,并且用release出来的v2 10-320 model做pre-train(其实不做pre_train loss 基本也都是nan),一开始只有clc_loss_score0和clc_loss_score1有value,其他loss均为nan, 没有更改过其他代码,data也是用的release出来的widerface pkl文件,请问这问题该怎么解决呀?
from lffd-a-light-and-fast-face-detector-for-edge-devices.
@Wenstery 那这个现象很奇怪,你能把config的设置给我看一眼么.
from lffd-a-light-and-fast-face-detector-for-edge-devices.
@YonghaoHe 感谢回复!现在重新核对了一些和原始文件的改动,training loss正常了
from lffd-a-light-and-fast-face-detector-for-edge-devices.
@Wenstery 你好,请问我在loss回传的时候也遇到了跟你一样的问题,在cls_loss回传中,会出现Nan的现象而且bbox loss下降的比较快,请问遇到这个问题你是怎么解决的呢?
from lffd-a-light-and-fast-face-detector-for-edge-devices.
@QingL0218 @Wenstery PyTorch 版本的LFD已经发布了,详情请关注我的主页,谢谢。
from lffd-a-light-and-fast-face-detector-for-edge-devices.
Related Issues (20)
- 头部侦测问题 HOT 13
- Four point detection HOT 4
- 头部检测问题 HOT 6
- Nan loss while training face detection model on custom dataset HOT 2
- How to run inside Nvidia Deepstream?
- face detection model failed in the CPU environment HOT 3
- Error in face detection code while testing over WIDER_val dataset
- Error in validation over WIDER_val
- Difference on accuracy between your results and my reproduction HOT 4
- LFFD使用给的pytorch模型无法测试出论文的wider结果
- Please update the repo for Vehicle Detection HOT 6
- 人脸检测结果不符合预期
- 请问运行效率fps是在什么数据集上得到的
- Not realtime when run face_detection program on GPU
- Detecting Tiny Faces HOT 3
- face_detect-python-opencv-dnn
- Basketball tracking and Intel Movidius Myriad X support
- Cuda 10.2+TensorRT 7.0.0.1, Inference issue, ValueError: could not broadcast input array from shape (3,4,3) into shape (3,4,1)
- Slowly when run 2 cameras on 1 machine
- AttributeError: module 'mxnet' has no attribute 'module'
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 lffd-a-light-and-fast-face-detector-for-edge-devices.