Datasets:

srinathnr
/

TTS_DATASET

Tasks:

Modalities:

Text

Languages:

English

Size:

10K<n<100K

License:

Dataset card Data Studio Files Files and versions Community

Dataset Preview

Split (3)

train

The full dataset viewer is not available (click to read why). Only showing a preview of the rows.

The computed response content exceeds the supported size in bytes (10000000).

Error code:   TooBigContentError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

text string
h# p l iː z h# k ɔː l h# s t ɛ l ə h#
h# æ s k h# h ɜː h# t ə h# b ɹ ɪ ŋ h# ð iː z h# θ ɪ ŋ z h# w ɪ ð h# h ɜː h# f ɹ ʌ m ð ə h# s t ɔːɹ h#
h# s ɪ k s h# s p uː n z h# ʌ v h# f ɹ ɛ ʃ h# s n oʊ h# p iː z h# f aɪ v h# θ ɪ k h# s l æ b z h# ʌ v h# b l uː h# tʃ iː z h# æ n d h# m eɪ b iː h# ɐ h# s n æ k h# f ɔːɹ h# h ɜː h# b ɹ ʌ ð ɚ h# b ɑː b h#
h# w iː h# ɔː l s oʊ h# n iː d h# ɐ h# s m ɔː l h# p l æ s t ɪ k h# s n eɪ k h# æ n d h# ɐ h# b ɪ ɡ h# t ɔɪ h# f ɹ ɑː ɡ h# f ɚ ð ə h# k ɪ d z h#
h# ʃ iː h# k æ n h# s k uː p h# ð iː z h# θ ɪ ŋ z h# ɪ n t ʊ h# θ ɹ iː h# ɹ ɛ d h# b æ ɡ z h# æ n d h# w iː h# w ɪ l h# ɡ oʊ h# m iː t h# h ɜː h# w ɛ n z d eɪ h# æ t h# ð ə h# t ɹ eɪ n h# s t eɪ ʃ ə n h#
h# w ɛ n h# ð ə h# s ʌ n l aɪ t h# s t ɹ aɪ k s h# ɹ eɪ n d ɹ ɑː p s h# ɪ n ð ɪ h# ɛɹ h# ð eɪ h# æ k t h# æ z h# ɐ h# p ɹ ɪ z ə m h# æ n d h# f ɔːɹ m h# ɐ h# ɹ eɪ n b oʊ h#
h# ð ə h# ɹ eɪ n b oʊ h# ɪ z h# ɐ h# d ᵻ v ɪ ʒ ə n h# ʌ v h# w aɪ t h# l aɪ t h# ɪ n t ʊ h# m ɛ n i h# b j uː ɾ i f əl h# k ʌ l ɚ z h#
h# ð iː z h# t eɪ k h# ð ə h# ʃ eɪ p h# ə v ə h# l ɔ ŋ h# ɹ aʊ n d h# ɑːɹ tʃ h# w ɪ ð h# ɪ t s h# p æ θ h# h aɪ h# ə b ʌ v h# æ n d h# ɪ t s h# t uː h# ɛ n d z h# ɐ p æ ɹ ə n t l i h# b ᵻ j ɔ n d h# ð ə h# h ɚ ɹ aɪ z ə n h#
h# ð ɛ ɹ h# ɪ z h# ɐ k ɔːɹ d ɪ ŋ h# t ə h# l ɛ dʒ ə n d h# ɐ h# b ɔɪ l ɪ ŋ h# p ɑː t h# ʌ v h# ɡ oʊ l d h# æ t h# w ʌ n h# ɛ n d h#
h# p iː p əl h# l ʊ k h# b ʌ t h# n oʊ w ʌ n h# ɛ v ɚ h# f aɪ n d z h# ɪ t h#
h# w ɛ n h# ɐ h# m æ n h# l ʊ k s h# f ɔːɹ h# s ʌ m θ ɪ ŋ h# b ᵻ j ɔ n d h# h ɪ z h# ɹ iː tʃ h# h ɪ z h# f ɹ ɛ n d z h# s eɪ h# h iː h# ɪ z h# l ʊ k ɪ ŋ h# f ɚ ð ə h# p ɑː t h# ʌ v h# ɡ oʊ l d h# æ t h# ð ɪ h# ɛ n d h# ʌ v ð ə h# ɹ eɪ n b oʊ h#
h# θ ɹ uː aʊ t h# ð ə h# s ɛ n tʃ ɚ ɹ i z h# p iː p əl h# h æ v h# ɛ k s p l eɪ n d h# ð ə h# ɹ eɪ n b oʊ h# ɪ n h# v ɛ ɹ iə s h# w eɪ z h#
h# s ʌ m h# h æ v h# ɐ k s ɛ p t ᵻ d h# ɪ ɾ h# æ z h# ɐ h# m ɪ ɹ ə k əl h# w ɪ ð aʊ t h# f ɪ z ɪ k əl h# ɛ k s p l ɐ n eɪ ʃ ə n h#
h# t ə h# ð ə h# h iː b ɹ uː z h# ɪ t h# w ʌ z ɐ h# t oʊ k ə n h# ð æ t h# ð ɛɹ h# w ʊ d h# b iː h# n oʊ m ɔːɹ h# j uː n ɪ v ɜː s əl h# f l ʌ d z h#
h# ð ə h# n ɔːɹ s ɪ m ɛ n h# k ə n s ɪ d ɚ d h# ð ə h# ɹ eɪ n b oʊ h# æ z h# ɐ h# b ɹ ɪ dʒ h# oʊ v ɚ h# w ɪ tʃ h# ð ə h# ɡ ɑː d z h# p æ s t h# f ɹ ʌ m h# ɜː θ h# t ə h# ð ɛɹ h# h oʊ m h# ɪ n ð ə h# s k aɪ h#
h# ʌ ð ɚ z h# h æ v h# t ɹ aɪ d h# t ʊ h# ɛ k s p l eɪ n h# ð ə h# f ɪ n ɑː m ɪ n ə n h# f ɪ z ɪ k l i h#
h# æ ɹ ɪ s t ɑː ɾ əl h# θ ɔː t h# ð æ t ð ə h# ɹ eɪ n b oʊ h# w ʌ z h# k ɔː z d h# b aɪ h# ɹ ᵻ f l ɛ k ʃ ə n h# ʌ v ð ə h# s ʌ n z h# ɹ eɪ z h# b aɪ h# ð ə h# ɹ eɪ n h#
h# s ɪ n s h# ð ɛ n h# f ɪ z ɪ s ɪ s t s h# h æ v h# f aʊ n d h# ð ɐ ɾ ɪ t h# ɪ z h# n ɑː t h# ɹ ᵻ f l ɛ k ʃ ə n h# b ʌ t h# ɹ ᵻ f ɹ æ k ʃ ə n h# b aɪ h# ð ə h# ɹ eɪ n d ɹ ɑː p s h# w ɪ tʃ h# k ɔː z ᵻ z h# ð ə h# ɹ eɪ n b oʊ z h#
h# m ɛ n i h# k ɑː m p l ᵻ k eɪ ɾ ᵻ d h# aɪ d iə z h# ɐ b aʊ t h# ð ə h# ɹ eɪ n b oʊ h# h ɐ v b ɪ n h# f ɔːɹ m d h#
h# ð ə h# d ɪ f ɹ ə n s h# ɪ n ð ə h# ɹ eɪ n b oʊ h# d ᵻ p ɛ n d z h# k ə n s ɪ d ɚ ɹ ə b l i h# ə p ɑː n h# ð ə h# s aɪ z h# ʌ v ð ə h# d ɹ ɑː p s h# æ n d h# ð ə h# w ɪ t θ h# ʌ v ð ə h# k ʌ l ɚ d h# b æ n d h# ɪ ŋ k ɹ iː s ᵻ z h# æ z h# ð ə h# s aɪ z h# ʌ v ð ə h# d ɹ ɑː p s h# ɪ ŋ k ɹ iː s ᵻ z h#
h# ð ɪ h# æ k tʃ uː əl h# p ɹ aɪ m ɚ ɹ i h# ɹ eɪ n b oʊ h# ə b z ɜː v d h# ɪ z h# s ɛ d h# t ə b i h# ð ɪ h# ɪ f ɛ k t h# ʌ v h# s uː p ɚ ɹ ɪ m p ə z ɪ ʃ ə n h# ə v ə h# n ʌ m b ɚ ɹ h# ʌ v h# b oʊ z h#
h# ɪ f h# ð ə h# ɹ ɛ d h# ʌ v ð ə h# s ɛ k ə n d h# b oʊ h# f ɔː l z h# ə p ɑː n h# ð ə h# ɡ ɹ iː n h# ʌ v ð ə h# f ɜː s t h# ð ə h# ɹ ɪ z ʌ l t h# ɪ z h# t ə h# ɡ ɪ v h# ɐ h# b oʊ h# w ɪ ð h# ɐ n h# ɐ b n ɔːɹ m əl i h# w aɪ d h# j ɛ l oʊ h# b æ n d h# s ɪ n s h# ɹ ɛ d h# æ n d h# ɡ ɹ iː n h# l aɪ t h# w ɛ n h# m ɪ k s t h# f ɔːɹ m h# j ɛ l oʊ h#
h# ð ɪ s h# ɪ z h# ɐ h# v ɛ ɹ i h# k ɑː m ə n h# t aɪ p h# ʌ v h# b oʊ h# w ʌ n h# ʃ oʊ ɪ ŋ h# m eɪ n l i h# ɹ ɛ d h# æ n d h# j ɛ l oʊ h# w ɪ ð h# l ɪ ɾ əl h# ɔːɹ h# n oʊ h# ɡ ɹ iː n h# ɔːɹ h# b l uː h#
h# ð eɪ h# ɑːɹ h# v ɛ ɹ i h# k l ɛ v ɚ ɹ h# ɪ n h# ð ɛɹ h# j uː s h# ʌ v h# w ɜː d z h#
h# m iː ɾ ɪ ŋ z h# w ɪ l h# ɔː l s oʊ h# ɹ ᵻ m eɪ n h# p ɹ aɪ v ə t h#
h# s ə dʒ ɛ s tʃ ə n z h# ʌ v ð ɪ h# æ k ʃ ə n h# b iː ɪ ŋ h# ɐ h# p ʌ n ɪ ʃ m ə n t h# w ɜː h# d ɪ s m ɪ s t h#
h# b ɪ k ʌ z h# w iː h# d uː n ɑː t h# n iː d h# ɪ t h#
h# ɐ h# ʃ ɔːɹ t l ɪ s t h# ɪ z h# t ə b i h# d ɹ ɔː n h# ʌ p h# n ɛ k s t h# m ʌ n θ h#
h# n ʌ n ð ə l ɛ s h# ð ɪ h# oʊ v ɚ ɹ ɔː l h# p ɪ k tʃ ɚ ɹ h# ɪ z h# h ɛ l θ i h#
h# aɪ h# f iː l h# s oʊ h# d ɛ s p ɚ ɹ ə t l i h# s ɑː ɹ i h# f ɔːɹ h# ɡ ɔːɹ d ə n h# æ n d h# s ɛ ɹ ə h#
h# ɪ ɾ h# ɪ z h# n ɑː t h# f ɔː ɹ h# ɛ n i h# ʌ ð ɚ h# p ɜː p ə s h#
h# ɪ t h# k æ ɹ i d h# ɐ h# d ʌ b əl h# m ɛ s ɪ dʒ h#
h# ð ə h# k ʌ m p ə n i h# d ɪ d n ɑː t h# d ᵻ k l ɛ ɹ h# ɐ h# d ɪ v ɪ d ɛ n d h#
h# ð æ t h# ɪ z h# ð ə h# f eɪ s h# ʌ v h# f ɪɹ h#
h# ʌ v h# k ɔːɹ s h# ɔ n h# t uː z d eɪ h# j uː n aɪ ɾ ᵻ d h# w ɜː h# b iː ʔ n̩ h# d ᵻ s p aɪ t h# ð ɪ s h#
h# ð ɪ h# æ m b ɪ ʃ ə n h# ð ɪ s h# j ɪ ɹ h# ɪ z h# t ʊ h# ɪ m p ɹ uː v h# ɔ n h# l æ s t h#
h# j uː h# m ʌ s t h# ɔː l w eɪ z h# ɐ t ɛ m p t h# t ə h# ɹ eɪ z h# ð ə h# b ɑːɹ h#
h# ɪ t h# d ᵻ p ɛ n d z h# ɔ n ð ə h# d ᵻ s ɪ ʒ ə n z h# ʌ v ð ə h# m ɛ m b ɚ h# s t eɪ t s h#
h# ɪ ɾ h# ɪ z h# s ɪ ɹ i ə s h# b ʌ t h# n ɑː t h# ð æ t h# s ɪ ɹ i ə s h#
h# ð ə h# b ɔɪ h# ɪ z h# n oʊ h# l ɑː ŋ ɡ ɚ h# w ɪ ð h# ʌ s h#
h# ð eɪ h# s ɛ d h# ʃ iː h# w ʌ z h# s t eɪ b əl h#
h# ɐ m ʌ ŋ h# ð ɛ m h# ɑːɹ h# m ɛ n i h# k ɹ ɪ m ɪ n əl z h#
h# m ɜː dʒ ɚ h# w ʊ d h# b iː h# ɛ n t aɪɚ l i h# l ɑː dʒ ɪ k əl h#
h# ɛ v ɹ ɪ θ ɪ ŋ h# ɪ z h# p ɑː s ᵻ b əl h# ɪ n h# f ʊ t b ɔː l h#
h# ɪ t s h# n ɑː ɾ ə h# s w ɪ m ɪ ŋ h# p uː l h# aʊ t h# ð ɛɹ h#
h# j ʊɹ ɹ ə p h# w ɪ l h# b ɛɹ h# ð ə h# b ɹ ʌ n t h# ʌ v ð ə h# k ʌ t s h#
h# ɪ t h# h æ d h# l ɪ ɾ əl h# ɪ f ɛ k t h# ɔ n h# j ʊɹ ɹ ə p iə n h# ɛ k w ᵻ ɾ i h# m ɑːɹ k ɪ t s h#
h# m eɪ b iː h# f ʊ l t aɪ m h# ɹ ɛ f ɚ ɹ iː z h# w ɪ l h# p ɹ ə v aɪ d h# ð ɪ h# æ n s ɚ h#
h# ð ɪ h# aʊ t k ʌ m h# ʌ v h# ð ɛɹ h# ɡ eɪ m h# ɪ z ə n t h# ð ɪ h# ɪ m p ɔːɹ t ə n t h# θ ɪ ŋ h#
h# w iː h# w ʊ d h ɐ v h# k ə n s ɜː n z h# ɐ b aʊ t h# s ə s p ɛ n ʃ ə n h#
h# ð ə h# k ʌ p əl h# æ k tʃ uː əl i h# k eɪ m h# ə p ɑː n h# ð ɪ h# æ k s ɪ d ə n t h#
h# ð ə h# p ɹ ɑː b l ə m z h# ɑː ɹ h# ɐ h# ɹ ɪ z ʌ l t h# ʌ v h# ð æ t h# ʃ ɔːɹ t f ɔː l h#
h# ɪ ɾ ə l i h# ɪ z h# ɐ h# b j uː ɾ i f əl h# p l eɪ s h#
h# w iː h# h æ v h# t ə h# k iː p h# aʊ ɚ h# k l aɪə n t s h#
h# ð ə h# ɹ eɪ t h# ʌ v h# ɡ ɹ oʊ θ h# ɪ n h# ɹ oʊ d h# t ɹ æ f ɪ k h# ɪ z h# ɔː l ɹ ɛ d i h# b ɪ ɡ ɪ n ɪ ŋ h# t ə h# s l oʊ h#
h# w aɪ h# d uː h# j uː h# w ɔ n t h# t ə h# k ʌ m h# t ʊ h# ɛ d ɪ n b ʌ ɹ ə h#
h# h aʊ h# ɑːɹ h# j uː h# s ɜː h#
h# s oʊ h# ɪ z h# ð ɐ ɾ ɪ t h# ð ɛ n h#
h# d uː h# aɪ h# h æ v h# ɐ h# f eɪ v ɹ ɪ t h#
h# w ʌ t h# d uː h# w iː h# d uː h#
h# w aɪ h# d uː h# ɪ ɾ h# ʌ ð ɚ w aɪ z h#
h# w ʌ t h# k aɪ n d h# ʌ v h# m æ n h# d ʌ z h# ð æ t h# m ɪ s t ɚ h# d ɪ k h#
h# dʒ ɔːɹ dʒ h# ɹ ɑː b ɚ t s ə n h# d ᵻ s k ɹ aɪ b d h# ð ə h# p l æ n h# æ z h# ɐ n h# aʊ t ɹ eɪ dʒ ə s h# æ n d h# d ɪ s ɡ ɹ eɪ s f əl h# d ᵻ s ɪ ʒ ə n h#
h# w iː h# w ɪ l h# p eɪ h# ð ɛɹ h# b ɪ l z h#
h# w iː h# k ʊ d h# s uː n h# b iː h# ɪ n h# ð æ t h# s eɪ m h# p ə z ɪ ʃ ə n h#
h# ɪ t h# k eɪ m h# æ z h# ɐ h# s ɚ p ɹ aɪ z h#
h# ð ə h# f ɑːɹ m ɚ h# w ɔ n t s h# ɐ h# n uː h# k æ m p eɪ n h#
h# oʊ n l i h# w ɔː ɾ ɚ ɹ h# ɪ z h# k ɜː ɹ ə n t l i h# ɐ l aʊ d h#
h# ɪ t s h# n ɑː t h# k ʌ v ɚ d h# b aɪ h# ð ɪ h# ɪ n ʃ ʊɹ ɹ ə n s h#
h# f aɪ n əl h# p ɹ ə p oʊ z əl z h# w ɪ l h# b iː h# p ʌ b l ɪ ʃ t h# ɪ n h# f ɛ b ɹ uː ɛ ɹ i h#
h# j uː h# ʃ ʊ d h# t ɹ aɪ h# t ə h# h æ v h# ɐ h# l æ f h#
h# ɹ ɑː b ɪ n h# w ɪ l j ə m z h# ɪ z h# v ɛ ɹ i h# s ʌ b d uː d h#
h# ɪ t h# w ʌ z n̩ t h# p ɑː s ᵻ b əl h# t ə h# ɹ ᵻ l æ k s h#
h# aɪ h# w ʌ z h# v ɛ ɹ i h# p l iː z d h# w ɪ ð h# ð ɛ m h# t ə d eɪ h# ð oʊ h#
h# ð ə h# s p ɪ ɹ ɪ t h# ɪ n ð ə h# d ɹ ɛ s ɪ ŋ h# ɹ uː m h# ɪ z h# t ɹ ə m ɛ n d ə s h#
h# aɪ h# w ʌ z h# æ z h# s t ɹ ɔ ŋ h# æ z h# ɐ h# h ɔːɹ s h#
h# ð eɪ h# d uː n ɑː t h# w ɜː k h# f ɔːɹ h# ɡ l æ z ɡ oʊ h# s ɪ ɾ i h# k aʊ n s əl h#
h# b oʊ θ h# m ɛ n h# w ɜː ɹ h# ɔ n h# w ʌ n j ɪɹ h# k ɑː n t ɹ æ k t s h#
h# h æ l ɪ f æ k s h# h ɐ z h# ɔː l s oʊ h# b ɪ n h# m ɛ n ʃ ə n d h# æ z h# ɐ h# l aɪ k l i h# p ɹ ɛ d ə ɾ ɚ h#
h# ʃ iː h# d aɪ d h# ɪ n h# h ɑː s p ɪ ɾ əl h# t uː h# aʊ ɚ z h# l eɪ ɾ ɚ h#
h# ɪ t h# h æ p ə n z h# ɔ n h# ɐ n h# æ n j uː əl h# b eɪ s ɪ s h#
h# ɪ ɾ h# ɪ z h# ɐ h# dʒ ɑː b h# k ɹ iː eɪ ʃ ə n h# s k iː m h#
h# ð ə h# p ɑːɹ ɾ i h# ɪ z h# ʌ p h# f ɔː ɹ h# ɪ t h#
h# iː v ə n h# ð ə h# w ʌ n h# ʃ iː h# l ʌ v d h#
h# ɪ t h# m eɪ k s h# f ɚ ɹ ə n h# ɪ n t ɹ ɛ s t ɪ ŋ h# k ɑː n f ɹ ə n s h#
h# aɪ v h æ d h# ɪ t h# f ɚ ð ɪ h# ɛ ɡ z æ m z h#
h# ð oʊ z h# p iː p əl h# m ʌ s t h# b iː h# ɐ k aʊ n t ə b əl h#
h# w iː h# f iː l h# v ɛ ɹ i h# k ʌ m f t ə b əl h# ɪ n h# ð ɪ s h# ɪ n t ɚ n æ ʃ ə n əl h# ɛ n v aɪ ɹ ə n m ə n t h#
h# p ɛ t ɹ əl h# p ɹ aɪ s ᵻ z h# w ɜː h# w aɪ d l i h# b l eɪ m d h#
h# w iː h# h æ v h# t ʊ h# ɔ f ɚ h# v æ l j uː h# f ɔːɹ h# m ʌ n i h#
h# f l ɔː ɹ ɪ d ə h# ɪ z h# ð ə h# p ɪ v ə ɾ əl h# s t eɪ t h# ɪ n ð ə h# n eɪ ʃ ə n h#
h# ɔ n h# s æ ɾ ɚ d eɪ h# ɪ t h# w ʌ z h# m ʌ tʃ h# ð ə h# s eɪ m h#
h# ð eɪ h# l ɪ v d h# f ɔːɹ h# ð ɛɹ h# tʃ ɪ l d ɹ ə n h#
h# p oʊ ʃ ə n z h# s ɛ d h# h iː h# w ʊ d h# d ɹ aɪ v h#
h# ð ə h# w ʊ m ə n h# ɪ z h# b eɪ s ɪ k l i h# ɔ f h# h ɜː h# h ɛ d h#
h# aɪ h# k ʊ d ə n t h# ɹ iə l i h# w ɜː k h# aʊ t s aɪ d h# ɛ d ɪ n b ʌ ɹ ə h#
h# h iː h# h æ d h# m ɪ s t h# ɐ h# p ɛ n əl ɾ i h# ɐ ɡ ɛ n s t h# ð ɪ h# ɪ ŋ ɡ l ɪ ʃ h#
h# ð ɪ h# ɔɪ l h# k ʌ m p ə n i z h# w ʊ d h# b iː h# æ s k t h# t ə h# k ə n t ɹ ɪ b j uː t h#
h# æ n d h# ð eɪ h# h æ d h# ɐ h# k eɪ s h#
h# ɪ t h# w ʌ z h# k l ɪ ɹ h# ɔ n h# θ ɜː z d eɪ h#

End of preview.

Phonemized-VCTK (speech + features)

Phonemized-VCTK is a light-repack of the VCTK corpus that bundles—per utterance—

the raw audio (wav/)
the plain transcript (txt/)
the IPA phoneme string (phonemized/)
frame-level pitch-aligned segments (segments/)
sentence-level context embeddings (context_embeddings/)
speaker-level embeddings (speaker_embeddings/)

The goal is to provide a turn-key dataset for
forced alignment, prosody modelling, TTS, and speaker adaptation experiments without having to regenerate these side-products every time.

Folder layout

Folder	Contents	Shape / format
`wav/<spk>/`	48 kHz 16‑bit mono `.wav` files	`p225_001.wav`, …
`txt/<spk>/`	original plain‑text transcript	`p225_001.txt`, …
`phonemized/<spk>/`	whitespace‑separated IPA symbols, `#h` = word boundary	`p225_001.txt`, …
`segments/<spk>/`	JSON with per‑phoneme timing & mean pitch	`p225_001.json`, …
`context_embeddings/<spk>/`	NumPy float32 `.npy`, sentence embedding of the utterance	`p225_001.npy`, …
`speaker_embeddings/`	NumPy float32 `.npy`, one vector per speaker, generated from NVIDIA `TitaNet-Large` model	`p225.npy`, …

Example segments entry

{
  "0": ["h#", {"start_sec":0.0,"end_sec":0.10,"duration_sec":0.10,"mean_pitch":0.0}],
  "1": ["p",  {"start_sec":0.10,"end_sec":0.18,"duration_sec":0.08,"mean_pitch":0.0}],
  "2": ["l",  {"start_sec":0.18,"end_sec":1.32,"duration_sec":1.14,"mean_pitch":1377.16}]
}

Quick start

from datasets import load_dataset

ds_train = load_dataset("srinathnr/TTS_DATASET", split="train", trust_remote_code=True, streaming=True)
ds_val = load_dataset("srinathnr/TTS_DATASET", split="validation", trust_remote_code=True, streaming=True)
ds_test = load_dataset("srinathnr/TTS_DATASET", split="test", trust_remote_code=True, streaming=True)

Custom Data Load

from pathlib import Path
from datasets import Audio
from torch.utils.data import Dataset

class CustomDataset(Dataset):
    def __init__(self, dataset_folder):
        self.dataset_folder = dataset_folder
        self.audio_files = sorted(
            [path for path in (Path(dataset_folder) / 'wav').rglob('*.wav') if not path.name.startswith('._')]
        )
        self.phoneme_files = sorted(
            [path for path in (Path(dataset_folder) / 'phonemized').rglob('*.txt') if not path.name.startswith('._')]
        )

        # Get the base file names (without extensions) for matching
        audio_basenames = {path.stem for path in self.audio_files}
        phoneme_basenames = {path.stem for path in self.phoneme_files}

        # Intersection of all file sets (excluding speaker embeddings)
        common_basenames = audio_basenames & phoneme_basenames

        # Filter files to only include common base names
        self.audio_files = [path for path in self.audio_files if path.stem in common_basenames]
        self.phoneme_files = [path for path in self.phoneme_files if path.stem in common_basenames]

        self.audio_feature = Audio(sampling_rate=16000)
    
    def __len__(self):
        return len(self.audio_files)

    def __getitem__(self, idx):
        audio_path = str(self.audio_files[idx])
        phoneme_path = str(self.phoneme_files[idx])
    
        align_audio = self.audio_feature.decode_example({"path": str(audio_path), "bytes": None})

        with open(phoneme_path, 'r') as f:
            phoneme = f.read()
        
        if phoneme is not None:
            phoneme = phoneme.split()
        else:
            phoneme = []

        return {
            'phoneme': phoneme,
            'align_audio': align_audio
        }

Explore

from pathlib import Path
import json, soundfile as sf
import numpy as np

root = Path("Phonemized-VCTK")

wav, sr = sf.read(root/"wav/p225/p225_001.wav")
text = (root/"txt/p225/p225_001.txt").read_text().strip()
ipa  = (root/"phonemized/p225/p225_001.txt").read_text().strip()
segs = json.loads((root/"segments/p225/p225_001.json").read_text())
ctx  = np.load(root/"context_embeddings/p225/p225_001.npy")

print(text)
print(ipa.split())       # IPA tokens
print(ctx.shape)         # (384,)

Known limitations

The phone set is plain IPA—no stress or intonation markers.
English only (≈109 speakers, various accents).
Pitch = 0 on unvoiced phones; interpolate if needed.
Embedding models were chosen for convenience—swap as you like.

Citation

Please cite both VCTK and this derivative if you use the corpus:

@misc{yours2025phonvctk,
  title        = {Phonemized-VCTK: An enriched version of VCTK with IPA, alignments and embeddings},
  author       = {Your Name},
  year         = {2025},
  howpublished = {\url{https://huggingface.co/datasets/your-handle/phonemized-vctk}}
}

@inproceedings{yamagishi2019cstr,
  title={The CSTR VCTK Corpus: English Multi-speaker Corpus for CSTR Voice Cloning Toolkit},
  author={Yamagishi, Junichi et al.},
  booktitle={Proc. LREC},
  year={2019}
}

Downloads last month: 1,413