STT Kab Conformer-Transducer Large

Model Overview

This collection contains large size version of Conformer-Transducer (around 120M parameters) trained on Mozilla Common Voice 10.0 Kabyle train set with around 131 hours of Kabyle speech from 141620 utterances and 145 unique speakers. The model transcribes speech in lower case latin alphabet along with spaces and apostrophes.

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 [2] was used for training the models for few epochs. These models are trained with this example script and this base config.

Datasets

Mozilla Common Voice (v10.0) Kabyle dataset

Tokenizer Construction

The tokenizer for this model was built using text corpus provided with the train dataset.

We build a Google Sentencepiece Tokenizer [3] with the following script:

python [NEMO_GIT_FOLDER]/scripts/tokenizers/process_asr_text_tokenizer.py \
  --manifest=dev_manifest.json,train_manifest.json \
  --vocab_size=1024 \
  --data_root="" \
  --tokenizer="spe" \
  --spe_type=bpe \
  --spe_character_coverage=1.0 \
  --spe_max_sentencepiece_length=4
  --log

Performance

The performance of Automatic Speech Recognition models is measured using Word Error Rate(WER) and Character Error Rate(CER).

The model obtains the following scores on the Mozilla Common Voice test set: 18.86 % WER, 6.74 % CER

How to Use this Model

The model is available for use in the NeMo toolkit [2], 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.EncDecRNNTBPEModel.from_pretrained(model_name="stt_kab_conformer_transducer_large")

Transcribing text with this model

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

Input

This model accepts 16000 Hz 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] NVIDIA NeMo Toolkit

[3] Google Sentencepiece Tokenizer

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.