import tensorflow as tf
import csv
new_model = tf.keras.models.load_model('model')

# Check its architecture
new_model.summary()

train_label = []
train_feature=[]
with open("test.csv","r") as r:
    reader = csv.reader(r)
    for i in reader:
        train_label.append((eval(i.pop())))
        b = [eval(j) for j in i]
        train_feature.append(b)
train_features = tf.constant(train_feature)
train_labels = tf.constant(train_label)

test_accuracy = tf.keras.metrics.Accuracy()

branch = 32
for num in range(len(train_features)//branch):
  # training=False is needed only if there are layers with different
  # behavior during training versus inference (e.g. Dropout).
  x = train_features[num*32:(num+1)*32]
  y = train_labels[num*32:(num+1)*32]
  logits = new_model(x, training=False)
  prediction = tf.argmax(logits, axis=1, output_type=tf.int32)
  print("真实值为",y,"预测值为",prediction)
  test_accuracy(prediction, y)

print("Test set accuracy: {:.3%}".format(test_accuracy.result()))