def predict(input_x):
out = ""
#if len(input_x) != (28*28):
#print("The input image or input array is shaped incorrectly. Expecting a 28x28 image.")
'''
for i in xrange(0,28):
out = out+"\n"
for( j in xrange(0,28)):
if (input_x[(i*28)+j]>0.5):
out = out + "1"
else:
out = out + "0"
'''
#print( "Input image array: \n", out)
E = evaluate_model()
#E.__init__()
prediction = E.evaluate_model(
model_version=20,
input=np.float32(input_x.reshape(28, 28)).flatten().reshape(
1, 784)) #FIRST ENTRY IS SAME AS ALL ENTRIES FOR CLASSIFIER
#print("\nPREDICTION CLASSIFIED:",prediction[0])
#prediction = int(random.random()*9.9) #Current prediction is random
return prediction[0]
def predict(input_x):
x = ""
array = []
array.append([])
if len(input_x) != (28 * 28):
print "The input image or input array is shaped incorrectly. Expecting a 28x28 image."
for i in xrange(0, 28):
x = x + "\n"
for j in xrange(0, 28):
if input_x[(i * 28) + j] > 0.5:
x = x + "1"
array[0].append(1)
else:
x = x + "0"
array[0].append(0)
print "Input image array: \n", x
#Fill in this function to correctly predict the class the image array corresponds to
#Fill in this function to correctly predict the class the image array corresponds to
#Fill in this function to correctly predict the class the image array corresponds to
#Fill in this function to correctly predict the class the image array corresponds to
prediction = emc.evaluate_model()
pre_result = prediction.evaluate_model(1, array)
return pre_result
import numpy as np
## makes sure these imports point to correct directory and correct files
import tf_classify_sensed_image as emc # evaluation file, takes one image in 28x28 and outputs class
import tf_train_model_ASSIGNMENT_FILE as tm # file we created in recitaiton
# get mnist as example
mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)
# model number
model = 1
## code calls the training file we created in recitation to train model
#model_train = tm.build_train() ### comment line out for evaluation only
#model_train.build_train_network(model) ### comment line out for evaluation only
# get single mnist image for evaluation
batch_xs, batch_ys = mnist.test.next_batch(1)
#print(batch_xs)
print(batch_ys)
# code opens saved trained model created by code above
model_eval = emc.evaluate_model()
image = model_eval.evaluate_model(model_version=model, input=batch_xs)
# prints out classification and correct label
print('Classification:')
print(image)
print('Correct Label:')
print(np.where(batch_ys[0] == 1))