AutoNN.CNN.cnn_generator.CNN

class CNN(in_channels:int,numClasses:int,config: Optional[list[tuple]]=None):

Parameters:

in_channels : The number of channels in an image
numClass : Total number of classes in a classification problem
config : Generated configuration by AutoNN's CreateCNN.create_config() method required to create a CNN model

Methods:

summary() :

Parameters:

input_shape : Shape of the image torch.tensor [C, H, W]
```
    where,  C = number of channels
            H = height of the image
            W = width of the image
```
border : bool datatype , default = True | It prints lines between layers while displaying the summary of a model if set to True
save()

classes : List of classes

image_shape : Tuple[int, int] (H,W) | dimension of the images

path : Optional | path to the directory where the model is intended to be stored

Note

For the very first time path should be included, otherwise it will throw an InvalidPathError exception.

filename : str | name of the model
load()

Parameters:

PATH : Path to the trained model.pth.

printmodel : bool datatype, default = False | print the model if set to True

loadmodel : bool datatype, default = True | loads all the stored weights if set to True
predict()

paths : Union[list | tuple] | list of unknown images for testing or prediction