Nnmnkwii

-------------------

- `49`_: Add support for build without cython
- `46`_: Cleanup frontnend implementation

-------------------

- `12`_: [experimental] Add :obj:`nnmnkwii.metrics` package
- `42`_: Fix installation failsure on no-utf-8 environments

-------------------

- `38`_: Add parameter trajectory smoothing.
- `37`_: Add ``tqdm`` as dependency. Dataset's ``asarray`` now report progress if ``verbose > 0``.
- `37`_: Add further support for incremental mean/var computation.
- `37`_: Add and improve normalization utilities, :func:`nnmnkwii.preprocessing.inv_scale`, :func:`nnmnkwii.preprocessing.inv_minmax_scale` and :func:`nnmnkwii.preprocessing.minmax_scale_params`.
- Add builtin data source for Voice Conversion Challenge (VCC) 2016 dataset.
- `34`_: Add :func:`nnmnkwii.preprocessing.adjast_frame_length`.
- `34`_: ``adjast_frame_lengths`` now supports ``divisible_by`` parameter. ``ensure_even`` is deprecated.
- `34`_: Rename ``adjast_frame_length`` to ``adjast_Frame_lengths``
- Add references to :func:`nnmnkwii.postfilters.merlin_post_filter`.

-------------------

- `19`_: Achieved 80% test coverage
- `31`_: Cleanup data source implementations and add docs.
- Fix example data wasn't included in release tar ball.
- Support ``padded_length`` is ``None`` for :obj:`nnmnkwii.datasets.FileSourceDataset`.
- Automatic frame length adjastment for DTWAligner / IterativeDTWAligner

-------------------

- `28`_: Setuptools improvements. 1) __version__ now includes git commit hash. 2) description read README.rst using pandoc.
- `27`_: Add preemphasis / inv_preemphasis
- `26`_: Add tests for GMM based voice conversion if swap=True
- `25`_: fix typo in nnmnkwii/baseline/gmm.py

-------------------

- Add tests, achieve 75% test coverage.
- `23`_, `22`_: Preprocess rewrite & module restructure.
- `21`_: Add new function :obj:`nnmnkwii.autograd.UnitVarianceMLPG` that can run on CPU/GPU.