The number of peacocks present in a given audio clip is being identified, Model's accuracy has been improved using
1.Different structure of MLP having Multiple task Learning done.
2.Using Different loss functions and accuracy metrics.
3.Using class weights to account for imbalance.
4.Sampling techniques.
Results improved and are included, given more data corresponding to different classes this model would perform better than now.