CNN network AlexNet architecture explained with cat vs dog classifier example
Layer (type) | Output Shape | Param # |
---|---|---|
conv2d_1 (Conv2D) | (None, 55, 55, 96) | 34944 |
batch_normalization_1 | (Batch (None, 55, 55, 96) | 384 |
activation_1 (Activation) | (None, 55, 55, 96) | 0 |
max_pooling2d_1 | (None, 27, 27, 96) | 0 |
conv2d_2 (Conv2D) | (None, 27, 27, 256) | 614656 |
batch_normalization_2 | (None, 27, 27, 256) | 1024 |
activation_2 (Activation) | (None, 27, 27, 256) | 0 |
max_pooling2d_2 | (None, 13, 13, 256) | 0 |
conv2d_3 (Conv2D) | (None, 13, 13, 384) | 885120 |
conv2d_4 (Conv2D) | (None, 13, 13, 384) | 1327488 |
conv2d_5 (Conv2D) | (None, 13, 13, 256) | 884992 |
batch_normalization_3 | (None, 13, 13, 256) | 1024 |
activation_3 (Activation) | (None, 13, 13, 256) | 0 |
max_pooling2d_3 | (None, 6, 6, 256) | 0 |
flatten_1 (Flatten) | (None, 9216) | 0 |
dense_1 (Dense) | (None, 9216) | 84943872 |
batch_normalization_4 | (None, 9216) | 36864 |
activation_4 (Activation) | (None, 9216) | 0 |
dense_2 (Dense) | (None, 4096) | 37752832 |
batch_normalization_5 | (None, 4096) | 16384 |
activation_5 (Activation) | (None, 4096) | 0 |
dense_3 (Dense) | (None, 1) | 4097 |
batch_normalization_6 | (None, 1) | 4 |
activation_6 (Activation) | (None, 1) | 0 |
This model is built with AlexNet architecture on around 25000 Cat and Dog images. This model has a accuracy more than 72% on test set.
This dataset is downloaded from kaggle weblsite. The data set can be downloaded from the link given in this repository file. This data set contains 2 classes Cat and Dog. This data set contain total of 25000 examples (including training ,validation and testing sets).