This collection contains medium size versions of Conformer-Transducer (around 32M parameters) trained on NeMo ASRSet with over 16000 hours of english speech. The model transcribes speech in lower case english alphabet along with spaces and apostrophes.
Conformer-Transducer model is an autoregressive variant of Conformer model  for Automatic Speech Recognition which uses Transducer loss/decoding. You may find more info on the detail of this model here: Conformer-Transducer Model.
The tokenizers for these models were built using the text transcripts of the train set with this script.
All the models in this collection are trained on a composite dataset (NeMo ASRSET) comprising of several thousand hours of English speech:
Note: older versions of the model may have trained on smaller set of datasets.
The list of the available models in this collection is shown in the following table. Performances of the ASR models are reported in terms of Word Error Rate (WER%) with greedy decoding.
|Version||Tokenizer||Vocabulary Size||LS test-other||LS test-clean||WSJ Eval92||WSJ Dev93||NSC Part 1||MLS Test||MLS Dev||MCV Test 6.1||Train Dataset|
|1.6.0||SentencePiece Unigram||1024||4.7||2.1||1.8||2.6||6.3||7.2||6.4||9.3||NeMo ASRSET 2.0|
|1.4.0||SentencePiece Unigram||1024||5.8||2.6||2.4||3.2||7.8||-||-||-||NeMo ASRSET|
The model is available for use in the NeMo toolkit , and can be used as a pre-trained checkpoint for inference or for fine-tuning on another dataset.
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.EncDecRNNTBPEModel.from_pretrained(model_name="stt_en_conformer_transducer_medium")
python [NEMO_GIT_FOLDER]/examples/asr/transcribe_speech.py \ pretrained_name="stt_en_conformer_transducer_medium" \ audio_dir="PATH_TO_AUDIO_FILES"
This model accepts 16000 KHz Mono-channel Audio (wav files) as input.
This model provides transcribed speech as a string for a given audio sample.
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.