pydrobert

pydrobert is the namespace I use for the python utilities that I redistribute open source. Subpackage XXX can be found on PyPI with the name pydrobert-XXX and on Conda with the same name on the channel sdrobert:

pip install pydrobert-XXX
conda install -c sdrobert pydrobert-XXX
pip install git+https://github.com/sdrobert/pydrobert-XXX  # bleeding edge

The package can then be imported via

import pydrobert.XXX

This is student-driven code, so don't expect a stable API. I'll try to use semantic versioning, but the best way to keep functionality stable is by pinning the version in the requirements or by forking.

Available Packages

• pydrobert-kaldi: Pythonic bindings for Kaldi I/O. Unlike other packages, pydrobert-kaldi needs no external Kaldi installation to work. Compiled for Windows, OSX, and Linux.
• pydrobert-param: Intuitive hyperparameter (de)serialization that couples with the param package. Includes hooks to read parameters in from a file automatically using argparse arguments.
• pydrobert-pytorch: PyTorch modules, utilities, and various bits-and-bobs I use when building models for PyTorch. Includes gradient estimators, edit-distance-based metrics, data loaders, experiment management, attention mechanisms, and other things.
• pydrobert-speech: Digital signal processing for speech. This includes the standard Mel-scaled filter banks, and so much more. The results can be saved to NumPy, PyTorch, or Kaldi.

Defunct Packages

• pydrobert-gpyopt: Wrappers to make GPyOpt more accessible. PyPI only. Dumped in favour of Optuna.

Licensing

All subpackages are Apache 2.0 licensed.

How To Cite

If you use any pydrobert subpackage in an academic publication, please cite as follows:

@misc{robertsonPydrobert2019,
  title = {pydrobert},
  url = {https://github.com/sdrobert/pydrobert},
  author = {Robertson, Sean},
  year = {2019}
}