Table 2.

Different convolutional neural network architectures.

CNNs	Parameters (feature extractor layers)	Final model size (mb)	Images processed per second	Source
HawksheadNet	20,256	0.32–13.5	~6114.18	In-house CNN
EfficientNetB0	4,049,564	17.0	~ 393.57	https://www.kaggle.com/models/tensorflow/efficientnet/TensorFlow2/b0-feature-vector/1
EfficientNetB1	6,575,232	27.2	~300.04	https://www.kaggle.com/models/tensorflow/efficientnet/TensorFlow2/b1-feature-vector/1
EfficientNetB2	7,768,562	32	~260.03	https://www.kaggle.com/models/tensorflow/efficientnet/TensorFlow2/b2-feature-vector/1
EfficientNetB3	10,783,528	44.2	~207.55	https://www.kaggle.com/models/tensorflow/efficientnet/TensorFlow2/b3-feature-vector/1
EfficientNetB4	17,673,816	72–72.4	~162.91	https://www.kaggle.com/models/tensorflow/efficientnet/TensorFlow2/b4-feature-vector/1
EfficientNetB5	28,513,520	115.5	~113.87	https://www.kaggle.com/models/tensorflow/efficientnet/TensorFlow2/b5-feature-vector/1
EfficientNetB6	40,960,136	165.5	~81.37	https://www.kaggle.com/models/tensorflow/efficientnet/TensorFlow2/b6-feature-vector/1
EfficientNetB7	64,097,680	258.3	~67.45	https://www.kaggle.com/models/tensorflow/efficientnet/TensorFlow2/b7-feature-vector/1
InceptionV1	5,607,184	23.1	~572.10	https://www.kaggle.com/models/google/inception-v1/TensorFlow2/feature-vector/2
InceptionV2	10,173,112	41.4	~449.58	https://www.kaggle.com/models/google/inception-v2/TensorFlow2/feature-vector/2
InceptionV3	21,802,784	88.4	~379.57	https://www.kaggle.com/models/google/inception-v3/TensorFlow2/feature-vector/2
MobileNetV1 (1.0 × depth)	3,228,864	13.5	~811.90	https://www.kaggle.com/models/google/mobilenet-v1/TensorFlow2/100-224-feature-vector/2
MobileNetV2 (1.0 × depth)	2,257,984	9.8	~678.21	https://www.kaggle.com/models/google/mobilenet-v2/TensorFlow2/100-224-feature-vector/2
MobileNetV3-Large (1.0 × depth)	1,529,968	17.7	~629.59	https://www.kaggle.com/models/google/mobilenet-v3/TensorFlow2/large-100-224-feature-vector/1
MobileNetV3-Small (1.0 × depth)	4,226,432	6.7	~1251.15	https://www.kaggle.com/models/google/mobilenet-v3/TensorFlow2/small-100-224-feature-vector/1