pydrobert-pytorch

PyTorch utilities for Machine Learning. This is an eclectic mix of utilities that I’ve used in my various projects. There is a definite leaning towards speech, specifically end-to-end ASR. The primary benefit pydrobert-pytorch has over other packages is modularity: you can pick and choose the functionality you desire without subscribing to an entire ecosystem. You can find out more about what the package offers in the documentation links below.

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.

Documentation

• Latest

• v0.3.0

Installation

pydrobert-pytorch is available through both Conda and PyPI.

conda install -c sdrobert pydrobert-pytorch
pip install pydrobert-pytorch

Licensing and How to Cite

Please see the pydrobert page for more details on how to cite this package.

Implementations of pydrobert.torch._img.{polyharmonic_spline,sparse_image_warp} are based off Tensorflow’s codebase, which is Apache 2.0 licensed.

Implementations of pydrobert.torch._compat.{broadcast_shapes,TorchVersion,one_hot} were directly taken from the PyTorch codebase. A number of methods and functions in pydrobert.torch._straight_through modify PyTorch code (see the file for more info). PyTorch has a BSD-style license which can be found in the file LICENSE_pytorch.txt.

The implementation of pydrobert.torch._compat.check_methods was taken directly from the CPython codebase, Copyright 2007 Google with additional notices at https://docs.python.org/3/copyright.html?highlight=copyright.

The file pydrobert.torch._textgrid,py was taken with some minor modifications from nltk_contrib. It is Apache 2.0-licensed, with the specific license text saved to LICENSE_nltk.txt.

Tutorials

• Attention and Transformer Networks
• Language Modelling and Decoding
  • A Simple-ish Example
  • Extensions

API

• pydrobert.torch
  • argcheck
  • config
  • data
  • distributions
  • estimators
  • functional
  • lightning
  • modules
  • training

Other

• Command-Line Interface
  • chunk-torch-spect-data-dir
  • compute-mvn-stats-for-torch-feat-data-dir
  • compute-torch-token-data-dir-error-rates
  • ctm-to-torch-token-data-dir
  • get-torch-spect-data-dir-info
  • print-torch-ali-data-dir-length-moments
  • print-torch-ref-data-dir-length-moments
  • subset-torch-spect-data-dir
  • textgrids-to-torch-token-data-dir
  • torch-ali-data-dir-to-torch-token-data-dir
  • torch-spect-data-dir-to-wds
  • torch-token-data-dir-to-ctm
  • torch-token-data-dir-to-textgrids
  • torch-token-data-dir-to-torch-ali-data-dir
  • torch-token-data-dir-to-trn
  • trn-to-torch-token-data-dir
• References

Indices and tables

• Index

• Module Index

• Search Page