Model Overview

Citrinet-512 model which has been trained on the open source Aishell-2 Mandarin Chinese corpus.

It utilizes a character encoding scheme, and transcribes text in the standard character set that is provided in the Aishell-2 Mandard Corpus.

Model Architecture

Citrinet is a deep residual convolutional neural network architecture that is optimized for Automatic Speech Recognition tasks. There are many variants of the Citrinet family of models, which are further discussed in the paper [1].

Training

This model was initially trained on the roughly 7,000 hours of speech compiled from various public English speech corpus, then fine-tuned on the open source Aishell-2 [2] corpus consisting of about 1000 hours transcribed Mandarin speech. The NeMo toolkit [3] was used for training this model over several hundred epochs on multiple GPUs.

Datasets

While training this model, we used the following datasets:

• Librispeech 960 hours of English speech (pre-training)
• Fisher Corpus (pre-training)
• Switchboard-1 Dataset (pre-training)
• WSJ-0 and WSJ-1 (pre-training)
• National Speech Corpus - 1 (pre-training)
• Mozilla Common Voice (pre-training)
• Aishell-2 corpus (fine-tuning)

Performance

The performance of Automatic Speech Recognition models is measuring using Character Error Rate. Since this dataset is pre-trained on a much larger speech corpus, and fine-tuned on this dataset, it will generally perform better at transcribing audio.

The model obtains the following scores on the following evaluation datasets -

• 6.3 % on Aishell-2 dev_ios
• 6.4 % on Aishell-2 test_ios

Note that these scores on Aishell-2 are not particularly indicative of the quality of transcriptions that models trained on ASR Set will achieve, but they are a useful proxy.

How to Use this Model

The model is available for use in the NeMo toolkit [4], and can be used as a pre-trained checkpoint for inference or for fine-tuning on another dataset.

Automatically load the model from NGC

import nemo.collections.asr as nemo_asr
asr_model = nemo_asr.models.EncDecCTCModel.from_pretrained(model_name="stt_zh_citrinet_512")

Transcribing text with this model

python [NEMO_GIT_FOLDER]/examples/asr/transcribe_speech.py \
  pretrained_name="stt_zh_citrinet_512" \
  audio_dir=""

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.

Limitations

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

References

[1] Citrinet: Closing the Gap between Non-Autoregressive and Autoregressive End-to-End Models for Automatic Speech Recognition

[2] AISHELL-2: Transforming Mandarin ASR Research Into Industrial Scale

[3] NVIDIA NeMo Toolkit

Licence

License to use this model is covered by the NGC TERMS OF USE unless another License/Terms Of Use/EULA is clearly specified. By downloading the public and release version of the model, you accept the terms and conditions of the NGC TERMS OF USE.