An easy implement of VGG19 with tensorflow, which has a detailed explanation.
The code is an implement of VGG19 with tensorflow. The detailed explanation can be found here.
Before running the code, you should confirm that you have :
- Python (2 and 3 is all ok, 2 need a little change on function"print()")
- tensorflow 1.0
- opencv
Then, you should download the model file "vgg19.npy" which can be found here or here(for users in china).
Finally, run the test file with "python3 testModel.py -m folder -p testModel", you will see some images with the predicted label (press any key to move on to the next image).
The command also supports url.
For eg. "python3 testModel.py -m url -p http://www.cats.org.uk/uploads/images/featurebox_sidebar_kids/Cat-Behaviour.jpg"
You can also use tensorboard to monitor the process. Remeber to see detailed explanation.
If you have any problem, please contact me!
blog ：http://blog.csdn.net/accepthjp
email ：huangjipengnju@gmail.com