Previously, when developer attempted to run the test process, they should Notes(#) some lines of code and add the code
test_image_array = ... # Better to be whitened in advance. Shape = [-1, img_height, img_width, img_depth]predictions = train.test(test_image_array) as mentioned in README.md.
It might be trouble to do these.
Thus, I add the test file named cifar10_test.py , with it, they can easily run the test prosess, just type in the script python cifar10_test.py. Afterwards, the prediction accuracy will be printed on the terminal, and a file named 'pridict_ret.txt' will be generated which contains the prediction log of the test process.
Based on my expariments, this cifar10_test.py works pretty well as shown below.
Previously, when developer attempted to run the test process, they should Notes(#) some lines of code and add the code
test_image_array = ... # Better to be whitened in advance. Shape = [-1, img_height, img_width, img_depth]predictions = train.test(test_image_array)as mentioned in README.md.It might be trouble to do these. Thus, I add the test file named cifar10_test.py , with it, they can easily run the test prosess, just type in the script
python cifar10_test.py. Afterwards, the prediction accuracy will be printed on the terminal, and a file named 'pridict_ret.txt' will be generated which contains the prediction log of the test process. Based on my expariments, this cifar10_test.py works pretty well as shown below.