Accelerating Inference and Language Model Fusion of Recurrent Neural Network Transducers via End-to-End 4-bit Quantization.
Andrea FasoliChia-Yu ChenMauricio J. SerranoSwagath VenkataramaniGeorge SaonXiaodong CuiBrian KingsburyKailash GopalakrishnanPublished in: CoRR (2022)
Keyphrases
- end to end
- language model
- recurrent neural networks
- language modeling
- n gram
- probabilistic model
- neural network
- speech recognition
- retrieval model
- document retrieval
- feed forward
- query expansion
- language modelling
- data fusion
- recurrent networks
- mixture model
- artificial neural networks
- test collection
- smoothing methods
- scalable video
- information retrieval
- bayesian inference
- context sensitive
- bayesian networks
- ad hoc information retrieval
- translation model
- fusion method
- congestion control
- relevance model
- image fusion
- information fusion
- query terms
- information retrieval systems
- multimedia