Transfer and Deep Learning-Based Gurmukhi Handwritten Word Classification Model

Table 5

Training parameters’ results of ResNet50 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	0.3299	0.3335	0.2503	0.2491
Validation accuracy	0.3929	0.4015	0.3031	0.3310
Training loss	2.1693	2.1658	2.4568	2.4608
Validation loss	1.9268	1.8842	2.2492	2.2037