hexiang10 / facial-expression-recognition Goto Github PK

利用Pytorch设计完成的基于卷积神经网络实现的面部表情识别项目 —— A facial expression recognition project based on convolution neural network designed by Pytorch 【Plus版本】：https://github.com/hexiang10/face-recognition-plus

Home Page: https://github.com/hexiang10/face-recognition-plus

Python 1.12% Jupyter Notebook 98.88%

python pytorch

facial-expression-recognition's Introduction

Hi there 👋 This is Hexiang ！✨

Welcome to my Github page! I am Hexiang 🙋‍♂️ Nice to meet you!

🌱 Things I am currently working on 🌱 :

Taking online courses about Software Engineering and Artificial Intelligence
My website 🚀 coming soon

⚡ Things recent ones I'm interested in ⚡ :

Chat GPT
Deep learning
Machine Learning
Cloud development
Cloud Service technology

💻 Programming languages and skills 💻 :

_{Self-assessment:
I am a person who loves life, sharing, and learning.I look forward to moving forward with you on the way to the future!}

If you want to know more, please chat to

facial-expression-recognition's People

Contributors

Stargazers

Watchers

facial-expression-recognition's Issues

路径问题

作者您好，我在训练vgg模型是为啥为出现FileNotFoundError: Couldn't find any class folder in face_images/vgg_verify_set.的错误
求解答，怎么设置才正确呢

Traceback (most recent call last):
File "F:\deeplearning\facial-expression-recognition\model_ResNet.py", line 32, in
data_vaild = torchvision.datasets.ImageFolder(root=path_vaild,transform=transforms_vaild)
File "F:\anaconda\envs\DL\lib\site-packages\torchvision\datasets\folder.py", line 310, in init
super().init(
File "F:\anaconda\envs\DL\lib\site-packages\torchvision\datasets\folder.py", line 145, in init
classes, class_to_idx = self.find_classes(self.root)
File "F:\anaconda\envs\DL\lib\site-packages\torchvision\datasets\folder.py", line 219, in find_classes
return find_classes(directory)
File "F:\anaconda\envs\DL\lib\site-packages\torchvision\datasets\folder.py", line 43, in find_classes
raise FileNotFoundError(f"Couldn't find any class folder in {directory}.")
FileNotFoundError: Couldn't find any class folder in face_images/vgg_verify_set.

Process finished with exit code 1

torchvision.datasets.ImageFolder错误

在您的VGG与ResNet代码中，使用了torchvision.datasets.ImageFolder来加载训练和验证集数据。这个函数假设数据按照文件夹进行组织，每个文件夹代表一个类别。然而，根据您的描述，您的数据集中只有一个文件夹，这可能导致了标签错误。如果您有测试过你的模型，可能会出现所有人脸都识别为序号0--即angry的表情，这并不是您所提到的“未训练完成”，而是“未正确训练”。而且训练过程中也会出现acc 100%的异常情况，我观察了你训练过程中的VGG与ResNet的演示图片，的确出现了这种异常。

在模型对比代码中，关于路径path_train 、path_vaild，在文档中说是放数据集和验证集，请问是train_set 和 verify_set吗？不是的话应该是哪两个路径？
我在跑模型对比代码时，老是出现错误ValueError: Found input variables with inconsistent numbers of samples: [4709, 9418]，请问与问题1有关吗？如果没关系请问是什么问题？

训练材料大小都不符

vgg和resnet要求训练材料是224224的，这个素材是4848的，楼主是怎么训练出来的啊

代码问题

文件是不少的，但是VGG和ResNet模型跑出来多次都结果一直是angry，检查多天代码也没找到问题，请问作者当时跑出来了吗