Model Overview

This collection contains large size versions of Conformer-Transducer (around 120M parameters) trained on 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 [2].

Model Architecture

Conformer-Transducer model is an autoregressive variant of Conformer model [1] 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 NeMo toolkit [3] was used for training the models. These model are trained with this example script and this base config. Some of the default parameters are different from the base config, and you may download the config file along with the nemo file.

Datasets

This model was trained on the roughly 1000 hours of speech from Aishell-2 [2].

Performance

The performance of Automatic Speech Recognition models is measuring using Character Error Rate.

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

• 5.0 % on Aishell-2 dev_ios
• 5.3 % on Aishell-2 test_ios
• 5.4 % on Aishell-2 dev_android
• 5.7 % on Aishell-2 test_android
• 5.5 % on Aishell-2 dev_mic
• 5.6 % on Aishell-2 test_mic

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_conformer_transducer_large")

Transcribing text with this model

python [NEMO_GIT_FOLDER]/examples/asr/transcribe_speech.py \
  pretrained_name="stt_zh_conformer_transducer_large" \
  audio_dir="<PATH_OF_AUDIO_FOLDER>"

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] Conformer: Convolution-augmented Transformer for 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.