Streaming English ASR

This page describes how to use sherpa for streaming ASR with Conformer transducer models trained with pruned stateless transdcuer.

Hint

To be specific, the pre-trained model is trained on the LibriSpeech dataset using the code from https://github.com/k2-fsa/icefall/tree/master/egs/librispeech/ASR/pruned_transducer_stateless4.

The pre-trained model can be downloaded from https://huggingface.co/pkufool/icefall_librispeech_streaming_pruned_transducer_stateless4_20220625

There are no recurrent modules in the transducer model:

  • The encoder network (i.e., the transcription network) is a Conformer model

  • The decoder network (i.e., the prediction network) is a stateless network, consisting of an nn.Embedding() and a nn.Conv1d().

  • The joiner network (i.e., the joint network) contains an adder, a tanh activation, and a nn.Linear().

Streaming ASR in this section consists of two components: