usef-kh / fer Goto Github PK

View Code? Open in Web Editor NEW

90.0 2.0 33.0 563 KB

Code for the paper "Facial Emotion Recognition: State of the Art Performance on FER2013"

Python 17.23% Jupyter Notebook 82.77%

saliency-maps optimizers architecture tuning loss vggnet fer2013 fer facial-emotions schedulers

fer's People

Contributors

Stargazers

Watchers

fer's Issues

The link of the trained model is broken

Hi!
I could not download the trained model from the link you provided. It seems it is broken.
Could you please fix the link?
link

How can I print the resulting probabilities of each class for each image?

output is nan

In utils > loops.py

outputs = net(inputs)

I check the output ,it's nan:

check = int((outputs != outputs).sum())
if(check>0):
print("your data contains Nan")
else:
print("Your data does not contain Nan, it might be other problem")

Then terminal print:

your data contains Nan
tensor([[nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan],
        [nan, nan, nan, nan, nan, nan, nan]], device='cuda:0',
       dtype=torch.float16, grad_fn=<AddmmBackward>)

I don't know how to solve this issues.

What is the reason why the test accuracy of 73% + is not achieved

Hello, I reproduce your project. I trained according to the readme file and obtained 92% + train accuracy and 73% + Eval accuracy.

Then I integrated the train and eval data as the train data.set train.py file has commented "scheduler = torch. Optim. Lr_scheduler. Cosiannealingwarmrestarts (optimizer, t_0 = 10, t_mult = 1, eta_min = 1e-6, last_epoch = - 1, verbose = true)" as the scheduler, and adjusted LR = 0.0001. After 50 epochs(start from 300epoch), only 72% + test accuracy was obtained.

What is the reason why the test accuracy of 73% + is not achieved

Error EOFError: Ran out of input

Hello.
Manage work experience
But it gives me this error

What is the solution?

Is it possible to acces the model ? The link is broken

About the accuracy

Hi I have tested your model for 300 epochs without editing any codes. But I cannot achieve the same accuracy of 73.28%. I wanna ask for your GPU edition to try again in the same environment. Thanks a lot!

Pretrained model question

Hi @usef-kh , I got confused on certain point. Do you start training from scratch or you started from the VGG-16 pretrained model

Could you share the trained weights?

@usef-kh Hi Yousif, thanks for your valuable research, could you share the trained weights? I would appreciate it if you could do it.

Can't load checkpoint trained model 'VGGNet'

Hi, I'm jin
thank you for your sharing. it's very helpful for my project.

i have some problem to follow your research.

In 'code' section there are two files.
Evaluation.ipnyb and saliency map.ipnyb

i tried to follow the codes in Evaluation.ipnyb.

# Load Trained Model checkpoint = torch.load('VGGNet') net = Vgg().to(device) net.load_state_dict(checkpoint["params"]) net.eval()

but, i got a error message such as
"FileNotFoundError: [Errno 2] No such file or directory: 'VGGNet'"

so i tried to find trained model "VGGNet"

Then, i thougt VGGNet = vgg.py
so, i applied the directory for use vgg.py.
just like,
checkpoint = torch.load('/Users/jincho/Desktop/study/techeer_bootcamp/fer-master/models/vgg.py')

but there was another error like "UnpicklingError: could not find MARK"

May i ask the right answer to fix these problems?

Sincerely,

How can we visualize trained model using Grad-CAM?

Hey @usef-kh,

It's really great work.

I have implemented this repo with no errors, great documentation and code. I just want to visualize my results with trained model using Grad-CAM. Can you guide me regarding visualizing results with trained model? any code or link to helpful material?

Thanks

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.