Transfer and Deep Learning-Based Gurmukhi Handwritten Word Classification Model

Table 4

Training parameters’ results of the proposed CNN model at different epochs and batch sizes.

Results	Training parameters
	Epoch = 100		Epoch = 40
	Batch size 20	Batch size 40	Batch size 20	Batch size 40
Training accuracy	0.9878	0.99	0.9486	0.9568
Validation accuracy	0.9973	0.9973	0.8985	0.9885
Training loss	0.0432	0.0313	0.1580	0.1440
Validation loss	0.0225	0.0188	1.0888	0.0404