Indic-TTS v1 Language Releases
Indic-TTS trains separate language systems using FastPitch acoustic models and HiFi-GAN V1 vocoders. Each monolingual system learns from male and female speakers.
Use the matching acoustic model and vocoder
The inference command needs both sets of weights and their configuration files. These releases differ from the later VITS and Parler families.AI4Bharat Indic-TTS project
Keep the paired components together
Practical guidance. Record the selected language archive, speaker setting and both model files. Replacing the vocoder changes the speech pipeline even when the text model is unchanged.
The release archive extends beyond the paper's language list
The paper reports thirteen Indian languages: Assamese, Bengali, Bodo, Gujarati, Hindi, Kannada, Malayalam, Manipuri, Marathi, Odia, Rajasthani, Tamil and Telugu.Towards Building Text-To-Speech Systems for the Next Billion Users
The release also includes en.zip, pa.zip and en+hi.zip. These archive names are preserved as English, Punjabi and English-Hindi variants. The release provides no separate training account for those three additions.Indic-TTS v1 checkpoint release
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | Downloadable |
| Model identifiers | Indic-TTS v1Indic-TTS AssameseIndic-TTS BengaliIndic-TTS BodoIndic-TTS GujaratiIndic-TTS HindiIndic-TTS KannadaIndic-TTS MalayalamIndic-TTS ManipuriIndic-TTS MarathiIndic-TTS OdiaIndic-TTS RajasthaniIndic-TTS TamilIndic-TTS TeluguIndic-TTS EnglishIndic-TTS PunjabiIndic-TTS English-Hindi |
| Architecture | FastPitch produces acoustic features; HiFi-GAN converts them to a waveform.AI4Bharat Indic-TTS project |
| License scope | The project repository uses MIT. The release page does not state separate checkpoint-specific terms.AI4Bharat Indic-TTS project |
| Published weights | Sixteen checkpoint ZIP archives.Indic-TTS v1 checkpoint release |
Related models
- Indic Parler-TTS
AI4Bharat · Indic Parler-TTS
Joint AI4Bharat and Hugging Face speech model controlled through written voice descriptions.
Launch date not confirmedRestricted accessDownloadable Speech-generation model for 11 Indian languages using text, reference audio and its transcript.
Launch date not confirmedRestricted accessDownloadable- VITS Rasa 13
AI4Bharat · Rasa TTS
Multilingual VITS speech model with predefined voices and styles for 13 Indian languages.
Launch date not confirmedRestricted accessDownloadable