A basic introduction to learning CNN through applications of VGG models.
#######################################################################################
This program demonstrates how to train a Convolutional Neural Network CNN or ConvNet
Firstly, we should download some photos as the dateset to train a VGG16 (a famous CNN model)
We download and create a flower datasets with 5 classes of totaling 3,670 images.
Secondly, we use the dataset to train a VGG16 model.
Thirdly, after having trained the VGG16 model, we predict an arbitrary flowers image.
The program will output the class name of the predicted image.
input dataset for training:
Download the "flower_photos.tgz" from below website, and put into "date_dir"
https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz
data_dir = "E:/datasets/flower_photos/"
input image for prediction: "673.jpg" or another images
output: print "roses" on the command prompt.
#######################################################################################
This program continues from the program, "vgg16_SaveModel.py". This program demonstrates how to load a CNN model and to predict a single image.
output(1): Save a model model.save("vgg16_t2.h5")
output(2): Save history of training save histroy into 'history_vgg16_t2.csv'
#######################################################################################
This program continues from the program, "vgg16_SaveModel.py". This program demonstrates how to load a CNN model and to predict a single image.
input(1): vgg16_t2.h5
input(2): image for prediction: "673.jpg" or another images
output: imshow the input image, and prediction and accuracy in %
#######################################################################################
This program continues from the program, "vgg16_SaveModel.py". This program demonstrates how to plot the training curves aftering training.
input: history_vgg16_t2.csv
Output: two charts
#######################################################################################
This program continues from the program, "vgg16_intro.py" and "vgg16_SaveModel.py".
This program demonstrates the architecture of vgg16 and vgg19
There are two main methods to construct a CNN model.
Method A: model = Sequential()
Method B: model = Model(inputs=input, outputs=output)
I prefer Method A for simple CNN. I will use Method B for more complicated CNN.
We also apply early stopping in training.
We can then use "vgg16_LoadModel.py" to load the saved model and predict a single image.
#######################################################################################
This program continues from the program, "vgg16_intro.py" and "vgg16_SaveModel.py".
This program demonstrates data augmentations, transfer learning and exponentialDecay learning rate.
We use both vgg16 and vgg19.
We also apply early stopping in training.
We can then use "vgg16_LoadModel.py" to load the saved model and predict a single image.