Operating environment:
anaconda4.4.0+python2.7.13
keras2.0.6+tensorflow1.0.1
GPU+CUDA+CUDNN
Discription:
This is my first time to use GitHub and write this passage,so if i write poorly,please take care of it thanks.
I have built a database which contain some pictures.These pictures is about bus,dinosaur,elephant,flower and horse,each of them have
hundreds of images.I have named five folders to put these images in.
Then,I use a function which called ImageDataGenerator.The usage of this Function can refer to here
How to use:
The project contain only two python file,'Train-with-cuda.py' and 'test.py'
'Train-with-cuda.py' is the file to train the neural network and save the model
'test.py' is the file to use the model we trained to classify images.
Attention:
The accuracy is not very high,it might be some parameter isn't the best.