Advances, Systems and Applications
Parameters | Details |
---|---|
CNN Model | Transfer Learning and Light-weight CNN (mobile-V3) with SVM |
No Training Layers | 100–250 |
Epochs | 50–100 |
Optimizer used | Stochastic Gradient Descent (SGD) Optimizer |
Model learning rate | 0.001 |
Loss function | Categorical Cross Entropy (CCE) function |
Activation function | SoftMax, ReLu |
Batch size | 256 |
Hidden Layer Architecture | (256,128,64) |