我试图创建一个以前在Imagenet上训练的InceptionV3 CNN。虽然检查点的创建和加载似乎工作正常,但结果似乎是随机的,因为每次运行脚本时,我都会得到不同的结果,即使我没有改变任何东西。网络是从头开始重新创建的,加载相同的未更改的网络并对相同的图像进行分类(根据我的理解,即使它无法确定图像实际是什么,仍然会导致相同的结果。)
我只是注意到,即使我尝试在同一个脚本执行中多次对同一图像进行分类,我也会得到随机结果。
我像这样使用
创建CNNfrom tensorflow.contrib.slim.nets import inception as nn_architecture
from tensorflow.contrib import slim
with slim.arg_scope([slim.conv2d, slim.fully_connected], normalizer_fn=slim.batch_norm,
normalizer_params={'updates_collections': None}): ## this is a fix for an issue where the model doesn't fit the checkpoint https://github.com/tensorflow/models/issues/2977
logits, endpoints = nn_architecture.inception_v3(input, # input
1001, #NUM_CLASSES, #num classes
# num classes #maybe set to 0 or none to ommit logit layer and return input for logit layer instead.
True, # is training (dropout = zero if false for eval
0.8, # dropout keep rate
16, # min depth
1.0, # depth multiplayer
layers_lib.softmax, # prediction function
True, # spatial squeeze
tf.AUTO_REUSE,
# reuse, use get variable to get variables directly... probably
'InceptionV3') # scope
之后我像这样加载the imagenet trained checkpoint
saver = tf.train.Saver()
saver.restore(sess, CHECKPOINT_PATH)
进行分类来验证它是否正常工作
我将它的原始分辨率压缩到299x299,这是网络的输入
from skimage import io
car = io.imread("data/car.jpg")
car_scaled = zoom(car, [299 / car.shape[0], 299 / car.shape[1], 1])
car_cnnable = np.array([car_scaled])
然后我尝试对图像进行分类,并尽可能打印图像所属的类别。
predictions = sess.run(logits, feed_dict={images: car_cnnable})
predictions = np.squeeze(predictions) #shape (1, 1001) to shape (1001)
print(np.argmax(predictions))
print(predictions[np.argmax(predictions)])
该类是(或似乎是)随机的,并且可能性也有所不同。 我的最后几次处决是:
Class - likelihood
899 - 0.98858
660 - 0.887204
734 - 0.904047
675 - 0.886952
以下是我的完整代码:https://gist.github.com/Syzygy2048/ddb8602652b547a71316ee0febfddbef
答案 0 :(得分:0)
由于我将isTraining设置为true,因此每次使用网络时都会应用辍学率。我的印象是这只发生在反向传播过程中。
要使其正常工作,代码应为
logits, endpoints = nn_architecture.inception_v3(input, # input
1001, #NUM_CLASSES, #num classes
# num classes #maybe set to 0 or none to ommit logit layer and return input for logit layer instead.
False, # is training (dropout = zero if false for eval
0.8, # dropout keep rate
16, # min depth
1.0, # depth multiplayer
layers_lib.softmax, # prediction function
True, # spatial squeeze
tf.AUTO_REUSE,
# reuse, use get variable to get variables directly... probably
'InceptionV3') # scope