Skip to content
This repository has been archived by the owner on Jun 13, 2024. It is now read-only.

Overloaded neutral? #11

Open
handav opened this issue Mar 19, 2019 · 2 comments
Open

Overloaded neutral? #11

handav opened this issue Mar 19, 2019 · 2 comments

Comments

@handav
Copy link

handav commented Mar 19, 2019

With the crossentropy model, performing softmax on the last dense layer, every image seems to have neutral as its highest (99%+) value. Is that in line with your results? Any suggestions for neutralizing neutral?

@amirhfarzaneh
Copy link

Do you mean that the network predicts every image in the test set as neutral?

@DanielXu123
Copy link

DanielXu123 commented Jan 8, 2020

I do have the same problem, I replace the FER 2013 emotion label with the mostly marked label in FER+, then trained that dataset with MobileNet. I get the model test accuracy with ~80% .

Then I randomly choose multi web downloaded images to run the forward progress. I get most results of neutral while some of the images should be happiness or sad

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants