This is the code for this video on Youtube by Siraj Raval as part of the Deep Learning Nanodegree course with Udacity. We're going to use an encoder-decoder model to generate a MIDI sequence given an input sequence of MIDI ragtime songs.
The program requires the following dependencies (easy to install using pip):
- Python 3
- TensorFlow (tested with v0.10.0rc0. Won't work with previous versions)
- CUDA (for using gpu, see TensorFlow installation page for more details)
- Numpy (should be installed with TensorFlow)
- Mido (midi library)
- Tqdm (for the nice progression bars)
- OpenCv (Sorry, there is no simple way to install it with python 3. It's primarily used as visualisation tool to print the piano roll so is quite optional. All OpenCv calls are contained inside the imgconnector file so if you want to use test the program without OpenCv, you can try removing the functions inside the file)
To train the model, simply run main.py. Once trained, you can generate the results with main.py --test --sample_length 500. For more help and options, use python main.py -h.
To visualize the computational graph and the cost with TensorBoard, run tensorboard --logdir save/.
Credits for the code go to Conchylicultor. I've merely created a wrapper to get people started.