Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Open source speech to text models for Indic Languages
| Date | Stars |
|---|---|
| 2026-07-24 | 327 |
| 2026-07-25 | 327 |
| 2026-07-28 | 327 |
| 2026-07-30 | 327 |
| 2026-08-06 | 327 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Vakyansh Open Source Models 1. [ Pretrained ASR Models ](#pretrained-asr-models) 2. [ Finetuned ASR Models ](#finetuned-asr-models) 3. [ Language Models ](#language-models) 4. [ Punctuation Models ](#punctuation-models) 5. [ TTS Models ](#tts-models) 6. [ Gender Classification Model ](#gender-classification-model) 7. [ Language Identification Models ](#language-identification-models) 8. [ Interspeech 2021 ASR Models ](#interspeech-2021-asr-models) <a name="pam"></a> ## Pretrained ASR Models **[wav2vec2-code](https://github.com/Open-Speech-EkStep/vakyansh-wav2vec2-experimentation)** | **[nemo-code](https://github.com/Open-Speech-EkStep/vakyansh-nemo-experimentation)** | Pretrained Model | Description | Architecture | Hours | |------------------|----------|----|---------| | [Vakyansh-Conformer-SSL](https://storage.googleapis.com/vakyansh-open-models/pretrained_models/vakyansh-conformer-ssl/ssl_conformer_large_e178.nemo) | This model was pre-trained using Nemo toolkit with 34,000 hours unlabeled audio in 39 Indian languages. This includes 15,000 hours of news recordings available on the internet, 10,000 hours of YouTube audios and other audio data. In addition, 9,000 hours of Indian English audio data was taken from NPTEL lectures open sourced by AI4Bharat. <br> This model was trained in collaboration with NVIDIA (NVIDIA Graphics Pvt Ltd). We thank NVIDIA for providing the compute resources to train this model. | Conformer-Large | 34,000 | | [CLSRIL-23](https://storage.googleapis.com/vakyansh-open-models/pretrained_models/clsril-23/CLSRIL-23.pt) | Cross Lingual Speech Representations for Indic Languages, Contains 10,000 hours of training data from 23 Indic Languages. <br> [Citation: https://arxiv.org/abs/2107.07402 ](https://arxiv.org/abs/2107.07402 ) | wav2vec2-Base | 10,000 | | [hindi_pretrained_4kh](https://storage.googleapis.com/vakyansh-open-models/pretrained_models/hindi/hindi_pretrained_4kh.pt) | Trained on 4200 hours of Hindi Data| wav2vec2-Base | 4,200 | | [kannada_pretrained_1400h](https://storage.googleapis.com/vakyansh-open-models/pretrained_models/kannada/kannada_pretrained_1400h.pt) | Trained on 1400 hours of Kannada data| wav2vec2-XLSR | 1,400 | <br><br> <a name="fam"></a> ## Finetuned ASR Models ### Conformer based models **[Repo](https://github.com/Open-Speech-EkStep/vakyansh-nemo-experimentation)** | Language | Pretrained Model | Finetuned Model | Finetuned Hours | Arch | |----|--------|----|-----|---| | Hindi | Vakyansh Conformer SSL | [hindi_large_ssl_2500](https://storage.googleapis.com/vakyansh-open-models/conformer_models/hindi/filtered_v1_ssl_2022-07-08_19-43-25/Conformer-CTC-BPE-Large.nemo) | 2,500 h | Large | | Indian English | Vakyansh Conformer SSL | [indian_en_large_ssl_700](https://storage.googleapis.com/vakyansh-open-models/conformer_models/english/2022-09-13_15-50-48/Conformer-CTC-BPE-Large.nemo) | 700 h | Large | | Kannada | Vakyansh Conformer SSL | [kannada_large_ssl_1000](https://storage.googleapis.com/vakyansh-open-models/conformer_models/kannada/iisc_noa_2022-08-30_22-45-15/Conformer-CTC-BPE-Large.nemo) | 1,000 h | Large | | Punjabi | Vakyansh Conformer SSL | [punjabi_large_ssl_500](https://storage.googleapis.com/vakyansh-open-models/conformer_models/punjabi/sme_noa_2022-08-23_19-56-08/Conformer-CTC-BPE-Large.nemo) | 500 h | Large | | Tamil | Vakyansh Conformer SSL | [tamil_large_ssl_900](https://storage.googleapis.com/vakyansh-open-models/conformer_models/tamil/iisc_noa_2022-09-03_14-06-50/Conformer-CTC-BPE-Large.nemo) | 900 h | Large | <br><hr> ### wav2vec2 based models **[Repo](https://github.com/Open-Speech-EkStep/vakyansh-wav2vec2-experimentation)** **Citation:** https://arxiv.org/abs/2203.16512 | Language | Pretrained Model | Finetuned Model | Dictionary | Single Model for Inference | Finetuned Hours | TS model | |----|--------|----|-----|-------------------|-------|---| | Hindi | CLSRIL-23 | [him_4200](https://storage.goog
Excerpt of 30,492 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:3edeca0a62b5d156, topic:speech-recognition, topic:stt, desc:speech-to-text