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 folder 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 url 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 ：[email protected]