Improvement in Automatic Speech Recognition of South Asian Accent Using Transfer Learning of DeepSpeech2

Table 4

Parameters and their values while implementation.

Parameters	Value
Operating system	Ubuntu 18.04
Frame work	Paddle paddle
Language	Python 3.7
CPU	Core-i7 (7th gen.)
RAM	16 GB (DDR3)
GPU	1080 Ti (11 GB memory, 3584 cores)
Batch size	256
Epochs	200
Drop out	0.4
Learning rate	l = 1 × e^−3
Loss function	Proposed Section 4.1.1
Optimizer	SGD