Stt Kr Conformer Transducer Large

🚀 stt_kr_conformer_transducer_large

This is a Conformer-Transducer model for Korean automatic speech recognition, offering high - quality transcription capabilities.

🚀 Quick Start

To quickly start using this model, you first need to install the necessary dependencies and then follow the usage examples.

✨ Features

• Automatic Speech Recognition: Specifically designed for Korean speech, it can accurately transcribe speech into text.
• Conformer - Transducer Architecture: Utilizes an autoregressive variant of the Conformer model with Transducer loss/decoding for better performance.
• Pre - trained and Fine - tuned: Based on a pre - trained English model and fine - tuned for Korean, it can be used as a pre - trained checkpoint for further fine - tuning or direct inference.

📦 Installation

To train, fine - tune or play with the model you will need to install NVIDIA NeMo. We recommend you install it after you've installed latest Pytorch version.

pip install nemo_toolkit['all']

💻 Usage Examples

Basic Usage

Automatically instantiate the model

import nemo.collections.asr as nemo_asr
asr_model = nemo_asr.models.ASRModel.from_pretrained("eesungkim/stt_kr_conformer_transducer_large")

Transcribing using Python

First, let's get a sample

wget https://dldata-public.s3.us-east-2.amazonaws.com/sample-kor.wav

Then simply do:

asr_model.transcribe(['sample-kor.wav'])

Advanced Usage

Transcribing many audio files

python [NEMO_GIT_FOLDER]/examples/asr/transcribe_speech.py  pretrained_name="eesungkim/stt_kr_conformer_transducer_large"  audio_dir="<DIRECTORY CONTAINING AUDIO FILES>"

Input

This model accepts 16000 KHz Mono - channel Audio (wav files) as input.

Output

This model provides transcribed speech as a string for a given audio sample.

📚 Documentation

Model Architecture

Conformer - Transducer model is an autoregressive variant of Conformer model [2] for Automatic Speech Recognition which uses Transducer loss/decoding. You may find more info on the detail of this model here: Conformer - Transducer Model.

Training

The model was finetuned based on the pre - trained English Model for over several epochs.

There are several transcribing and sub - word modeling methods for Korean speech recognition. This model uses sentencepiece subwords of Hangul characters based on phonetic transcription using Google Sentencepiece Tokenizer [3].

Datasets

All the models in this collection are trained on Ksponspeech dataset, which is an open - domain dialog corpus recorded by 2,000 native Korean speakers in a controlled and quiet environment. The standard split dataset consists of 965 hours of training set, 4 hours of development set, 3 hours of test - clean, and 4 hours of test - other.

Performance

Limitations

⚠️ Important Note

Since this model was trained on publically available speech datasets, the performance of this model might degrade for speech which including technical terms, or vernacular that the model has not been trained on. The model might also perform worse for accented speech.

💡 Usage Tip

This model produces a spoken - form token sequence. If you want to have a written form, you can consider applying inverse text normalization.

References

[1] NVIDIA NeMo Toolkit

[2] Conformer: Convolution - augmented Transformer for Speech Recognition

[3] Google Sentencepiece Tokenizer

📄 License

This model is licensed under CC - BY - 4.0.