Edit model card

Model Overview

NVIDIA NeMo: Training

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']

How to Use this Model

The model is available for use in the NeMo toolkit [1], and can be used as a pre-trained checkpoint for inference or for fine-tuning on another dataset.

Automatically instantiate the model

import nemo.collections.asr as nemo_asr
asr_model = nemo_asr.models.ASRModel.from_pretrained("ypluit/stt_kr_citrinet1024_PublicCallCenter_1000H")

Transcribing using Python

First, let's get a sample

get any korean telephone voice wave file

Then simply do:

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

Transcribing many audio files

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

Input

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

Output

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

Model Architecture

See nemo toolkit and reference papers.

Training

Learned about 10 days on 2 A6000

Datasets

Private call center real data (650hour)

Performance

0.13 CER

Limitations

This model was trained with 650 hours of Korean telephone voice data for customer service in a call center. might be Poor performance for general-purpose dialogue and specific accents.

References

[1] NVIDIA NeMo Toolkit

Downloads last month
5
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.