parler-tts/mls_eng
Viewer • Updated • 10.8M • 6.94k • 37
Open-source speech datasets annotated using Data-Speech
Open-source annotated speech datasets ranging from 1,000 hours to 45,000 hours.
Note The English version of the Multilingual LibriSpeech (MLS) dataset.
parler-tts/libritts_r_filtered
Viewer • Updated • 359k • 2.39k • 22Note Filtered version of the 1K high-quality LibriTTS-R dataset.
parler-tts/mls-eng-speaker-descriptions
Viewer • Updated • 10.8M • 93 • 11Note Annotations of English MLS above. Used for v1 training.
parler-tts/libritts-r-filtered-speaker-descriptions
Viewer • Updated • 359k • 70 • 8Note Annotations of the filtered LibriTTS-R dataset. Used for v1 training.
-
Parler-TTS
🥖848High-fidelity Text-To-Speech
-
Natural language guidance of high-fidelity text-to-speech with synthetic annotations
Paper • 2402.01912 • Published • 14
mythicinfinity/libritts_r
Viewer • Updated • 756k • 6.89k • 44Note A 1K hours high-quality English speech dataset.
parler-tts/mls_eng_10k
Viewer • Updated • 2.43M • 1.44k • 31Note A 10K hours subset of the English version of the Multilingual LibriSpeech (MLS) dataset.
parler-tts/mls-eng-10k-tags_tagged_10k_generated
Viewer • Updated • 2.43M • 163 • 17Note Annotations of the 10K hours subset of English MLS above. Used for v0.1 training.
parler-tts/libritts_r_tags_tagged_10k_generated
Viewer • Updated • 365k • 21 • 9Note An annotated version of LibriTTS-R above. Used for v0.1 training.
parler-tts/parler_tts_mini_v0.1
Text-to-Speech • 0.6B • Updated • 2.67k • 358Note A first model iteration of Parler-TTS, trained using the 10k hours of narrated audiobooks above.