Sub-8-bit quantization for on-device speech recognition: a regularization-free approach.
Kai ZhenMartin RadfarHieu Duy NguyenGrant P. StrimelNathan SusanjAthanasios MouchtarisPublished in: CoRR (2022)
Keyphrases
- speech recognition
- hidden markov models
- speech processing
- language model
- speech signal
- speech recognition technology
- speech understanding
- pattern recognition
- speech recognizer
- speech synthesis
- automatic speech recognition
- speech recognition systems
- speech recognizers
- noisy environments
- speaker dependent
- speaker identification
- cepstral coefficients
- speech recognition errors
- speech retrieval
- isolated word
- computer vision
- visual features
- speaker independent
- acoustic models
- feature vectors