search

marcogdepinto / emotion-classification-from-audio-files

Understanding emotions from audio files using neural networks and multiple datasets.
GNU General Public License v3.0
408 stars 134 forks source link

Overfitting except if you run it only in Ravdess_Music #12

Closed hellfireworld closed 4 years ago

hellfireworld commented 4 years ago

Very good work but I managed to obtain almost the same accuracy and loss only if I run the code in Ravdess_Music without Ravdess_Speech ! Else If I run it in both I get overfitting. The model has very good accuracy and loss and poor val_accuracy and val_loss

marcogdepinto commented 4 years ago

Hi, this model is built only to solve this problem, predicting emotions on the RAVDESS dataset. Overfitting is correct in this case because is what I wanted to achieve. If you want a more generalized model, feel free to reduce the number of layers in the neural network and/or remove/add different training data. This is not an issue but expected behaviour, hence I am resolving the issue.