---
license: cc-by-4.0
datasets:
- bene-ges/spellmapper_en_train
language:
- en
library_name: nemo
---
# SpellMapper - Spellchecking ASR Customization Model 

<style>
img {
 display: inline;
}
</style>

| [![Language](https://img.shields.io/badge/Language-en--US-lightgrey#model-badge)](#datasets)

This model is an alternative to word boosting/shallow fusion approaches:

- does not require retraining ASR model;
- does not require beam-search/language model (LM);
- can be applied on top of any English ASR model output;

Paper: [SpellMapper: A non-autoregressive neural spellchecker for ASR customization with candidate retrieval based on n-gram mappings](https://arxiv.org/abs/2306.02317)



## How to Use this Model

To use this model you will need to install [NVIDIA NeMo](https://github.com/NVIDIA/NeMo).

See [Bash-script](https://github.com/NVIDIA/NeMo/blob/main/examples/nlp/spellchecking_asr_customization/run_infer.sh) with example of inference pipeline.

Or play with [Tutorial](https://github.com/NVIDIA/NeMo/blob/stable/tutorials/nlp/SpellMapper_English_ASR_Customization.ipynb).


## Citation

```bibtex
    @misc{antonova2023spellmapper,
      title={SpellMapper: A non-autoregressive neural spellchecker for ASR customization with candidate retrieval based on n-gram mappings}, 
      author={Alexandra Antonova and Evelina Bakhturina and Boris Ginsburg},
      year={2023},
      eprint={2306.02317},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
    }
```