State Space Models

State Space Models#

Automatic Speech Recognition (ASR)#

asr-19m-v2-en-32b

A ~19m parameter state space ASR model trained on 15k hours of speech, achieving 10.61% average WER. See full details and download the model at huggingface.co/abr-ai/asr-19m-v2-en-32b.

Word error rates

Dataset

asr-19m-v2-en-32b

AMI-IHM

18.76%

Earnings-22

13.53%

GigaSpeech

15.44%

LibriSpeech (clean)

4.66%

LibriSpeech (other)

11.16%

SPGISpeech

3.94%

TED-LIUM

7.53%

VoxPopuli

9.88%

Average

10.61%