/tensorflow_cnn_tutorial

CSE480. Deep learning Class

Primary LanguageJupyter Notebook

tensorflow_cnn_tutorial

This repository is for practicing a tensorflow only

description

Answer.py: rename answer.py to Convolution.py if you want to apply answer.py

or

fill in Convolution.py like as answer.py(see models/answer.py). funtion run(), build_model(), demo() are little different from original

usage

To download:

$ ./download_and_preprocess_flowers.sh flowers

train:

$ python main.py

demo:

$ python main.py test/daisy.jpg

For demo, give the 'file_path' as a third argument. demo should be excuted after train.

Feedforward to CNN

Run feedforward network using mnist,

$ python mnist_tutorial.py

Run CNN using mnist,

$ python mnist_cnn_ver.py

This network has two convolutional layers and two fully-connected layers.