It is an attempt in the field of machine learning application on audio files . The data-set is taken from kaggle and the accuracy achieved is 83% on curated data-set and 70% on noisy data-set.
To better understand the problem statement you can visit this link :-Kaggle Comptetion.
To Download The Dataset use the link :- Dataset.
Many libraries while importing may give error . Please do install them and then try again .
The simple TKinter user interface is being used which is fully functional and takes the user input from the already saved files .
--> Text Files store the tags as per the order they need to be used in the model. --> The models and weights are separately stored and can be used as per user convenience. --> The TKinter User Interface is very basic but is dynamic as far as input is concerned.
If you want to just check the working you do not need to run the model files . Just use the Call or Tkinter files and they will give the specific tags.