There are some pre-trained models in Keras as follows:
- MobileNetV2: This model can be trained using ImageNet. The syntax if this model is as shown below:
keras.applications.mobilenet_v2.MobileNetV2 (
alpha = 1.0,
input_shape = None,
include_top = True,
input_tensor = None,
pooling = None,
weights = 'imagenet',
classes = 800
)
VGG16: This model is also trained using ImageNet and the syntax is shown below:
keras.applications.inception_v3.InceptionV3 (
include_top = True,
pooling = None,
classes = 2000
weights = 'imagenet',
input_tensor = None,
input_shape = None,
)
- InceptionV3:The syntax for this model is given below:
keras.applications.inception_v3.InceptionV3 (
include_top = True,
pooling = None,
classes = 2000
weights = 'imagenet',
input_tensor = None,
input_shape = None,
)
- InceptionResNetV2: This model can be loaded using the syntax below:
keras.applications.inception_resnet_v2.InceptionResNetV2 (
include_top = True,
input_tensor = None,
input_shape = None,
weights = 'imagenet',
pooling = None,
classes = 2000
)