STT Es FastConformer Hybrid Transducer-CTC Large P&C

Model Overview

This collection contains the Spanish FastConformer Hybrid (CTC and Transducer) Large model (around 114M parameters) with Punctuation and Capitalization. It is trained on the NeMo PnC ES ASRSET (Fisher, MCV12, MLS, Voxpopuli) containing 1424 hours of Spanish speech.

It utilizes a Google SentencePiece [1] tokenizer with vocabulary size 1024, and transcribes text in upper and lower case Spanish alphabet along with spaces, period, comma, question mark and inverted question mark.

Model Architecture

FastConformer is an optimized version of the Conformer model [2] with 8x depthwise-separable convolutional downsampling. The model is trained in a multitask setup with joint Transducer and CTC decoder loss. You may find more information on the details of FastConformer here: Fast-Conformer Model and about Hybrid Transducer-CTC training here: Hybrid Transducer-CTC.

Training

The NeMo toolkit [3] was used for training the models over several hundred epochs. These model are trained with this example script and this base config.

The tokenizers for these models were built using the text transcripts of the train set with this script.

Datasets

All the models in this collection are trained on a composite dataset (NeMo PnC ES ASRSET) comprising of 1424 thousand hours of Spanish speech:

• Fisher (141 hrs)
• MCV12 (395 hrs)
• MLS (780 hrs)
• Voxpopuli (108 hrs)

Tokenizer Construction

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

We build a Google Sentencepiece Tokenizer [1] with the following script :

python [NEMO_GIT_FOLDER]/scripts/tokenizers/process_asr_text_tokenizer.py \
  --manifest="train_manifest.json" \
  --data_root="<OUTPUT DIRECTORY FOR TOKENIZER>" \
  --vocab_size=1024 \
  --tokenizer="spe" \
  --spe_type="unigram" \
  --spe_character_coverage=1.0 \
  --no_lower_case \
  --log

Performance

The performance of Automatic Speech Recognition models is measuring using Word Error Rate. The model obtains the following scores on the following evaluation datasets.

a) On data without Punctuation and Capitalization using Transducer decoder for inference

b) On data with Punctuation and Capitalization using Transducer decoder for inference

How to Use this Model

The model is available for use in the NeMo toolkit [3], and can be used as a pre-trained checkpoint for inference or for fine-tuning on another dataset.

Automatically load the model from NGC

from nemo.collections.asr.models import ASRModel
asr_model = ASRModel.from_pretrained('stt_es_fastconformer_hybrid_large_pc')

Transcribing text with this model

Using Transuder decoder for inference:

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

Using CTC decoder for inference:

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

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. The model only outputs the punctuations: '.', ',', '¿', '?' and hence might not do well in scenarios where other punctuation is also expected.

References

[1] Google Sentencepiece Tokenizer

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

[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.