combnet 1.0.1

Combolutional Neural Networks

Combolutional Neural Networks

Training, evaluation, and implementations of Combolutional Layers in PyTorch

[Paper]

Table of contents

• Installation
• Training
  • Download
  • Preprocess
  • Partition
  • Train
  • Monitor
  • Evaluate
• References

Installation

You can install from PyPI: pip install combnet or from a local clone: pip install -e .

For full training and evaluation compatibility, you will also need to install FFMPEG version >=4, <7 (version 6 is recommended).

Training

Download

python -m combnet.data.download

Download and uncompress datasets used for training

Augmentation

python -m combnet.data.augment --datasets giantsteps_mtg

Augment data (pitch shift to other keys)

Preprocess

python -m combnet.data.preprocess --datasets giantsteps_mtg giantsteps

Preprocess datasets

Partition

python -m combnet.partition

Partition datasets. Partitions are saved in combnet/assets/partitions.

Train

python -m combnet.train --config <config> --gpus <gpus>

Trains a model according to a given configuration.

Monitor

Run tensorboard --logdir runs/. If you are running training remotely, you must create a SSH connection with port forwarding to view Tensorboard. This can be done with ssh -L 6006:localhost:6006 <user>@<server-ip-address>. Then, open localhost:6006 in your browser.

Evaluate

python -m combnet.evaluate \
    --config <config> \
    --checkpoint <checkpoint> \
    --gpu <gpu>

Evaluate a model. <checkpoint> is the checkpoint file to evaluate and <gpu> is the GPU index.