Skip to main content

Flowtron library

Project description

Flowtron library

This Flowtron library was changed to be used with vait library.

Original code, README.md and additional information:

https://github.com/NVIDIA/flowtron

Instalation

1) Install flowtron library

(This will also install waveglow library)

pip install flowtron==22.12.28

2) Install CUDA 11.3 or 11.6

pip install -r requirements-cuda-11.3.txt
# or
pip install -r requirements-cuda-11.6.txt

3) Install apex

git clone https://github.com/NVIDIA/apex /home/${USER}/apex
cd /home/${USER}/apex
pip install -v --disable-pip-version-check --no-cache-dir ./
cd -

4) Download published model Flowtron LJS

wget https://drive.google.com/open?id=1Cjd6dK_eFz6DE0PKXKgKxrzTUqzzUDW-

5) Download published model Flowtron LibriTTS

wget https://drive.google.com/open?id=1KhJcPawFgmfvwV7tQAOeC253rYstLrs8

6) Download published model Flowtron LibriTTS2K

wget https://drive.google.com/open?id=1sKTImKkU0Cmlhjc_OeUDLrOLIXvUPwnO

Usage: Training

flowtron-train -c config.json -p train_config.output_directory=outdir data_config.use_attn_prior=1

Usage: Inferencing

flowtron-inference -c config.json -f models/flowtron_ljs.pt -w models/waveglow_256channels_v4.pt -t "It is well know that deep generative models have a rich latent space!" -i 0

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

flowtron-22.12.28.tar.gz (38.4 kB view details)

Uploaded Source

Built Distribution

flowtron-22.12.28-py3-none-any.whl (44.7 kB view details)

Uploaded Python 3

File details

Details for the file flowtron-22.12.28.tar.gz.

File metadata

  • Download URL: flowtron-22.12.28.tar.gz
  • Upload date:
  • Size: 38.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.3.1 CPython/3.10.9 Linux/6.0.0-6-amd64

File hashes

Hashes for flowtron-22.12.28.tar.gz
Algorithm Hash digest
SHA256 38b2ac4c9cb5adee113f47f46920093a47953021e653190d485a558abac16cff
MD5 7320a9f35d7ae5b4ac26cdab81de088d
BLAKE2b-256 81d9fea957daba8e4b0a1943e1b817668458b6f8f44512a214e5c100d01d3028

See more details on using hashes here.

File details

Details for the file flowtron-22.12.28-py3-none-any.whl.

File metadata

  • Download URL: flowtron-22.12.28-py3-none-any.whl
  • Upload date:
  • Size: 44.7 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.3.1 CPython/3.10.9 Linux/6.0.0-6-amd64

File hashes

Hashes for flowtron-22.12.28-py3-none-any.whl
Algorithm Hash digest
SHA256 19b807d71dce004fd31927593040cdd833c21b4c8c2c37c90d6c6ef2eced1e45
MD5 54ea1105a4c1b3c98e9732e53b50600f
BLAKE2b-256 82a18814508742f7532f7d9f60a287646c09b02ae64e1d086b1561717c68aa75

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