Getting Started

Neural Network

A hidden size of 70 and 2 layers yield good results:

python train_hannds.py --hidden_size 70 --layers 2 --length 100 --cuda --cv_partition 1 --network 88 --rnn_type LSTM

Provide the --cuda flag to use the GPU.

python kalman_mapper.py

python hannds-server.py

Then open simple-text.maxpat in Max/MSP.

Name		Name	Last commit message	Last commit date
Latest commit History 92 Commits
data-hannds		data-hannds
pytest		pytest
LICENSE.txt		LICENSE.txt
README.md		README.md
evaluate_nn.py		evaluate_nn.py
evaluate_nn.sh		evaluate_nn.sh
hannds_data.py		hannds_data.py
hannds_files.py		hannds_files.py
hannds_server.py		hannds_server.py
kalman_mapper.py		kalman_mapper.py
network_zoo.py		network_zoo.py
requirements.yaml		requirements.yaml
rnntrainer.py		rnntrainer.py
simple-test.maxpat		simple-test.maxpat
train_hannds.py		train_hannds.py
train_midi_magenta.sh		train_midi_magenta.sh