bakingbrains/music_generation_using_deep_learning_-rnn-lstm-

Generating music using LSTM with the help of ESAC folk song dataset

Music_generation_using_deep_learning_-RNN-LSTM-

Generating music using LSTM with the help of ESAC folk song dataset

Dataset: https://kern.humdrum.org/cgi-bin/browse?l=essen%2Feuropa%2Fdeutschl

Libraries:

• music21 (pip install music21)
  
• tensorflow
  

Software required:

• musescore: https://musescore.org/en/download
  

After installing musescore, run the music21Environment.py with the correct path to musescore

• Download and copy the dataset to 'Dataset' folder
  
• The above provided file 'file_dataset' is the generated file from the preprocessing part for the folder named 'erk' from Dataset folder
  
• mapping.json is the json file for number of keys in the music file inside erk folder
  
• The trained model will be saved in log folder and used in musicgenerator.py file
  

File running sequence:

• music21Environment.py
• preprocessing.py
• train.py
• musicgenerator.py