VGG19 and VGG16 on Tensorflow
Switch branches/tags
Nothing to show
Clone or download
Latest commit dcc2449 Jun 12, 2017
machrisaa committed Jun 12, 2017 Merge pull request #35 from sdhnshu/master
Updated the ReadMe

Tensorflow VGG16 and VGG19

This is a Tensorflow implemention of VGG 16 and VGG 19 based on tensorflow-vgg16 and Caffe to Tensorflow. Original Caffe implementation can be found in here and here.

We have modified the implementation of tensorflow-vgg16 to use numpy loading instead of default tensorflow model loading in order to speed up the initialisation and reduce the overall memory usage. This implementation enable further modify the network, e.g. remove the FC layers, or increase the batch size.

To use the VGG networks, the npy files for VGG16 NPY or VGG19 NPY has to be downloaded.


Use this to build the VGG object

vgg = vgg19.Vgg19()


vgg = vgg16.Vgg16()

The images is a tensor with shape [None, 224, 224, 3].

Trick: the tensor can be a placeholder, a variable or even a constant.

All the VGG layers (tensors) can then be accessed using the vgg object. For example, vgg.conv1_1, vgg.conv1_2, vgg.pool5, vgg.prob, ... and contain the sample usage.


This library has been used in my another Tensorflow image style synethesis project: stylenet

Update 1: Trainable VGG:

Added a trainable version of the VGG19 vgg19_trainable. It support train from existing vaiables or from scratch. (But the trainer is not included)

A very simple testing is added test_vgg19_trainable, switch has demo about how to train, switch off train mode for verification, and how to save.

A seperated file is added (instead of changing existing one) because I want to keep the simplicity of the original VGG networks.

Update 2: Tensorflow v1.0.0:

All the source code has been upgraded to v1.0.0.

The conversion is done by my another project tf0to1