fvisionNetwork14 0.0.11

fvisionNetwork14 is an image classification CNN model that can classify the number of classes.

fvisionNetwork14

"fvisionNetwork14" is a CNN model for image classification that can categorize "n" classes. It has been tuned to have less codes than other models with higher code complexity. The model can categorize with improved accuracy with just a few lines of code. The dataset can be immediately fed into the model using an image pre-processing module that has been built into the package. Two graphical modules are given for plotting model accuracy and loss by providing model history as input.

Installation

pip install fvisionNetwork14

Released version

version: 0.0.11

Modules:

fvNet14

image_preprocessing

plot_accuracy

plot_loss

Pre-requisites:

tensorflow

Dependancy modules:

numpy

matplotlib

How to use?

Directory Structure

image_preprocessing

image_pre_processing.image_preprocessing(dataset_path, image_height = 50, image_width = 50)

Splitting train and test data using "image_array" and "class_label" from image_preprocessing module

X_train,x_test,Y_train,y_test = train_test_split(image_pre_processing.image_preprocessing.image_array,image_pre_processing.image_preprocessing.class_label,test_size=0.2,random_state=45)

fvNet14

deomo_model = fvNet14.fvNet14(image_height = 50, image_width = 50, color_channel = 3, output_layer = 10)
deomo_model.compile(loss= 'categorical_crossentropy', optimizer = 'adam', metrics = ['accuracy'])
history = deomo_model.fit(xtrain,ytrain,epochs=50,validation_data=(xtest,ytest))

plot_accuracy

plot_model_acuracy.plot_accuracy(history, height = 10, width = 10)

plot_loss

plot_model_loss.plot_loss(history, height = 10, width = 10)

License

© 2022 Kalyan Mohanty

This repository is licensed under the MIT license. See LICENSE for details.