Change log

v0.2.2 <2022-01-04>

  • #30: Fix typo in WaveNetTTS docs

  • #26: Fix accent phrase border in pp_symbols

  • #22: Fix wrong wavenet loss calculation (addresses #21)

  • #20: Fix: 毎回JSUTダウンロードをやり直す（全てのrun.shをFix）

  • #19: Enable Windows CI

  • #17: Add conv1d test to ensure forward/incremental_forward correctness

  • #14: windows: use expanduser instead of os.environ[“HOME”]

  • #13: Fix: 毎回JSUTダウンロードをやり直す

  • #12: Fix #10 書籍 p.82 code4.9 関数stftの誤植

  • #11: Add warning for streamlit online demo

v0.2.1 <2021-08-21>

  • pretrained: add PWG TTS models for common voice (ja)

  • pretrained: add HiFi-GAN based TTS models using JVS and JSUT corpus

  • Add HiFi-GAN configs for JVS and JSUT extra recipes

  • #7: Add script to generate ground-truth aligned (GTA) features

  • #5: [docker] Push docker image to Docker Hub

  • #4: [docker] fix docker build fail because no ‘gcc’ command

  • #2: [extra_recipes] Fix the suffix of the script; s/sh/bash/

  • #1: Add common voice jp recipes

v0.2.0 <2021-08-12>

The first public release!