Skip to main content

Aligner - PyTorch

Project description

Aligner - PyTorch

Sequence alignement methods with helpers for PyTorch.

Install

pip install aligner-pytorch

PyPI - Python Version

Usage

MAS

MAS (Monotonic Alignment Search) from GlowTTS. This can be used to get the alignment of any (similarity) matrix. Implementation in optimized Cython.

from aligner_pytorch import mas 

sim = torch.rand(1, 4, 6) # [batch_size, m_rows, n_cols]
alignment = mas(sim)

"""
sim = tensor([[
    [0.2, 0.8, 0.9, 0.9, 0.9, 0.4],
    [0.6, 0.8, 0.9, 0.7, 0.1, 0.4],
    [1.0, 0.4, 0.4, 0.2, 1.0, 0.7],
    [0.1, 0.3, 0.1, 0.7, 0.6, 0.9]
]])

alignment = tensor([[
    [1, 0, 0, 0, 0, 0],
    [0, 1, 1, 1, 0, 0],
    [0, 0, 0, 0, 1, 0],
    [0, 0, 0, 0, 0, 1]
]], dtype=torch.int32)
"""

Citations

Monotonic Alignment Search

@misc{2005.11129,
Author = {Jaehyeon Kim and Sungwon Kim and Jungil Kong and Sungroh Yoon},
Title = {Glow-TTS: A Generative Flow for Text-to-Speech via Monotonic Alignment Search},
Year = {2020},
Eprint = {arXiv:2005.11129},
}

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

aligner-pytorch-0.0.16.tar.gz (106.6 kB view details)

Uploaded Source

File details

Details for the file aligner-pytorch-0.0.16.tar.gz.

File metadata

  • Download URL: aligner-pytorch-0.0.16.tar.gz
  • Upload date:
  • Size: 106.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.1 CPython/3.9.13

File hashes

Hashes for aligner-pytorch-0.0.16.tar.gz
Algorithm Hash digest
SHA256 36b87badfbda97d44e715fa6c4a4b44149feaf90c75ee88b43eb9a0303559019
MD5 04c875cfce61d6f4c189c53f264ae2d5
BLAKE2b-256 ad6e6932ce00a041c71d2f52fc63d6d669ce4d71931e77e129a8d34d16fc16d6

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page