MiteshPuthran / Speech-Emotion-Analyzer

The neural network model is capable of detecting five different male/female emotions from audio speeches. (Deep Learning, NLP, Python)
MIT License
1.3k stars 438 forks source link

my val_accuracy only has 40+% #40

Closed bigcash closed 4 years ago

bigcash commented 4 years ago

i use the data, two of the RAVDESS and SAVEE. not 2000, but 1920。 i try to train a new model, but only got "942/942 [==============================] - 1s 2ms/step - loss: 0.2656 - accuracy: 0.9013 - val_loss: 1.9338 - val_accuracy: 0.4341" you can see that my train data has 942 item, not 1378. that's the problem.

MiteshPuthran commented 4 years ago

Try augmenting the data by adding more data files from different sources. It will improve the accuracy.