Sir, actually I am stuck on the vowel-consonant classification problem, Which is a subpart of the capstone project.
I am able to create my model, run it and even I achieved A good accuracy on the vowel classification part. Though, I am not able to tune the consonant separation hypermeters. If possible, can I share the my loss plots on both validation and test set?
Thanks.
Yeah sure, you can share them.
Sir, This is my above loss functions. The blur one, is the loss function obtained on the consonants. The Loss is decreasing, though the oscillation in both the curves are very high, which is causing overfitting in my model. Especially for the blue loss curve. I tried to lower the learning rate, but then it doesn't decrease. Remains constant. Can you please help on how to proceed after this.
Hi,
Which architecture and optimization algorithm are you using?
