---
tags:
- espnet
- automatic-speech-recognition
- speech
- audio
- indian-languages
language:
- bh
- bn
- ch
- hi
- kn
- mg
- mr
- mt
- te
license: cc-by-4.0
datasets:
- respin
pipeline_tag: automatic-speech-recognition
---

# ESPnet ASR Models in 9 Indian languages trained on RESPIN-S1.0 Small Train sets

This repository contains E-Branchformer-based ESPnet2 ASR models trained on the RESPIN-S1.0 small train splits.

## 🛠 Demo: How to use in ESPnet2

Follow the [ESPnet installation instructions](https://espnet.github.io/espnet/installation.html) if you haven't done that already.

```bash
cd espnet
pip install -e .
cd egs2/respin_small/asr1
./run.sh --skip_data_prep true --skip_train true --download_model SpireLab/spire_respin_baselines_espnet
```


## 📊 Results (CER/WER from RESULTS.md)

| Language | Model Name | CER (%) | WER (%) |
|----------|------------|---------|---------|
| bh | exp_small/exp_bh/asr_bh_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 4.4 | 15.2 |
| bn | exp_small/exp_bn/asr_bn_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 4.1 | 15.0 |
| ch | exp_small/exp_ch/asr_ch_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 3.1 | 10.6 |
| hi | exp_small/exp_hi/asr_hi_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 3.1 | 9.9 |
| kn | exp_small/exp_kn/asr_kn_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 4.6 | 24.5 |
| mg | exp_small/exp_mg/asr_mg_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 6.0 | 20.4 |
| mr | exp_small/exp_mr/asr_mr_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 3.1 | 14.5 |
| mt | exp_small/exp_mt/asr_mt_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 5.0 | 17.9 |
| te | exp_small/exp_te/asr_te_ebf_size256_mlp1024_lin1024_e8_mactrue_bs6M_gacc1 | 4.1 | 21.6 |

---

## 📚 Citation

```bibtex
@inproceedings{watanabe2018espnet,
  author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
  title={{ESPnet}: End-to-End Speech Processing Toolkit},
  year={2018},
  booktitle={Proc. Interspeech},
  pages={2207--2211},
  doi={10.21437/Interspeech.2018-1456},
  url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
}