Python 您必须使用 dtype float 为占位符张量“Placeholder”提供一个值
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/41607155/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
You must feed a value for placeholder tensor 'Placeholder' with dtype float
提问by judyzha
I'm a newer to tensorflow, I really don't know how to solve the problem.
我是 tensorflow 的新手，我真的不知道如何解决这个问题。
The code is like:
代码是这样的：
Feed the train with values:
sess.run(train_op, feed_dict={images: e, labels: l, keep_prob_fc2: 0.5})Use the value in CNN:
x = tf.placeholder(tf.float32, [None, 10 * 1024])
用值喂火车：
sess.run(train_op, feed_dict={images: e, labels: l, keep_prob_fc2: 0.5})使用 CNN 中的值：
x = tf.placeholder(tf.float32, [None, 10 * 1024])
Then have the error
然后有错误
InvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'Placeholder' with dtype float [[Node: Placeholder = Placeholder[dtype=DT_FLOAT, shape=[], _device="/job:localhost/replica:0/task:0/gpu:0"]()]]
InvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'Placeholder' with dtype float [[Node: Placeholder = Placeholder[dtype=DT_FLOAT, shape=[], _device="/job:localhost/replica:0/task:0/gpu:0"]()]]
I print the input valuetypes using print(e.dtype)and the result is float32and e.shape:(10, 32, 32, 1).
我使用打印输入值类型print(e.dtype)，结果是float32and e.shape:(10, 32, 32, 1)。
I really don't know why this error is happening.
我真的不知道为什么会发生这个错误。
The code format
代码格式
First:
第一的：
define the CNN model
"image = tf.placeholder(tf.float32, [FLAGS.batch_size, 32,32,1])" is here
Second:
第二：
loss funtion and train_op is here
"label = tf.placeholder(tf.float32, [None, FLAGS.batch_size])" is here
Third is the session:
三是座谈会：
images, labels = getShuffleimage()#here will get shuffle data
num_examples = 0
init = tf.initialize_local_variables()
with tf.Session() as sess:
# Start populating the filename queue.
sess.run(init)
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(coord=coord, sess=sess)
try:
step = 0
while not coord.should_stop():
start_time = time.time()
image, label = sess.run([images, labels])#get shuffle images
print(image.shape)
print(image.dtype)
sess.run(train_op, feed_dict={image: image, label: label , keep_prob_fc2: 0.5})
duration = time.time() - start_time
except tf.errors.OutOfRangeError:
print('Done training after reading all data')
finally:
# When done, ask the threads to stop.
coord.request_stop()
# Wait for threads to finish.
coord.join(threads)
sess.close()
采纳答案by xxi
Some questions
一些问题
first
why you use sess = tf.InteractiveSession()and with tf.Session() as sess:at same time, just curious
首先
你为什么使用sess = tf.InteractiveSession()，with tf.Session() as sess:同时，只是好奇
second
what is your placeholder name xor images?
if name is x, {images: x_data...}won't feed x_datato x, it override(?) images
I think feed_dict should be {x: x_data...}
第二你有什么占位符的名称x或images？
如果 name 是x，{images: x_data...}则不会提供x_data给x它，它会覆盖 (?)images
我认为 feed_dict 应该是{x: x_data...}
if name is images,do you have two imagesin your program, placeholderand shuffle data, try to modify name of variable
如果名称是images，images您的程序中是否有两个，placeholder以及shuffle data，尝试修改变量名称
回答by Targo
I saw one problem with the code. There are two variables with the same name label. One of them refers to a Tensor, and the other one refers to some data. When you set label: labelin the feed_dict, you need to distinguish between the two variables.
Maybe you can try changing the name for one of the variables?
我看到代码有一个问题。有两个同名的变量label。其中一个是指一个张量，另一个是指一些数据。在 中设置label: label时feed_dict，需要区分这两个变量。也许您可以尝试更改其中一个变量的名称？
