Bibliograph. Daten | Serai, Dhiren Devinder: Quantization of automatic speech recognition networks. Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik, Masterarbeit Nr. 84 (2020). 60 Seiten, englisch.
|
Kurzfassung | Recently, end-to-end neural networks based speech recognition systems have received great interests in the speech community. Although these systems offer a wide range of advantages such as high performance, they often rely on a neural network with a large number of parameters, i.e. large memory footprint and large decoding time. The main goal of this thesis is investigation of quantized neural networks for end-to-end speech recognition systems. Different quantization methods like post training quantization, scalar quantization, iterative product quantization and quantization aware training are investigated in this thesis on ASR model. By doing so, the final system is 3.7 times smaller than the baseline system, faster in decoding with a speedup factor of 1.19 and with almost same level of recognition performance.
|
Volltext und andere Links | Volltext
|
Abteilung(en) | Universität Stuttgart, Institut für Maschinelle Sprachverarbeitung
|
Betreuer | Vu, Prof. Ngoc Thang; Schweitzer, Dr. Antje; Denisov, Pavel |
Eingabedatum | 27. Juli 2021 |
---|