Wrong testing results on MNIIST data

[Azam Hamidinekoo]

Hello.

I have trained the LeNet on caffe on Mac. it was trained with 99% accuracy. Now I want to check it with some digits I downloaded from internet. 

I have written the code bellow.  

After executing this code, giving any numbers ( as images) I get [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.] for the score. Which shows something is wrong!!!!!

Could you please tell me what id wrong in testing section? I think the training was done completely!

Regards

--------------------------------- the code for testing)--------------------------------------

import numpy as np
import sys
import matplotlib.pyplot as plt
import os
import sys
caffe_root = '../'  # this file should be run from {caffe_root}/examples (otherwise change this line)
sys.path.insert(0, caffe_root + 'python')
import caffe

MODEL_FILE = '/a PATH/caffe-master/examples/mnist/lenet.prototxt'
PRETRAINED = '/a PATH/caffe-master/examples/mnist/lenet_iter_10000.caffemodel'
IMAGE_FILE = '/a PATH/2.png'       #image path

input_image = caffe.io.load_image(IMAGE_FILE, color=False)


net = caffe.Classifier(MODEL_FILE, PRETRAINED, image_dims=(28, 28), raw_scale=255)
score = net.predict([input_image], oversample=False)
print score