pixel-cnn "Object was never used" errors after upgrading to Tensorflow 1.2

After upgrading to Tensorflow 1.2, running train.py produces many (567) of the following error:

ERROR:tensorflow:==================================
Object was never used (type <class 'tensorflow.python.framework.ops.Tensor'>):
<tf.Tensor 'model_1/conv2d_0/stack:0' shape=(3,) dtype=int32>
If you want to mark it as used call its "mark_used()" method.
It was originally created here:
['File "train.py", line 135, in <module>\n    dropout_p=args.dropout_p, **model_opt)', 'File "/home/malcolm/data/anaconda3/lib/python3.6/site-packages/tensorflow/python/ops/template.py", line 261, in __call__\n    return self._call_func(args, kwargs, check_for_new_variables=True)', 'File "/home/malcolm/data/anaconda3/lib/python3.6/site-packages/tensorflow/python/ops/template.py", line 217, in _call_func\n    result = self._func(*args, **kwargs)', 'File "/media/bruce/MoreData/malcolm/tpc/base-pixel-cnn/pixel_cnn_pp/model.py", line 40, in model_spec\n    x_pad, num_filters=nr_filters, filter_size=[2, 3]))]  # stream for pixels above', 'File "/home/malcolm/data/anaconda3/lib/python3.6/site-packages/tensorflow/contrib/framework/python/ops/arg_scope.py", line 181, in func_with_args\n    return func(*args, **current_args)', 'File "/media/bruce/MoreData/malcolm/tpc/base-pixel-cnn/pixel_cnn_pp/nn.py", line 358, in down_shifted_conv2d\n    return conv2d(x, num_filters, filter_size=filter_size, pad=\'VALID\', stride=stride, **kwargs)', 'File "/home/malcolm/data/anaconda3/lib/python3.6/site-packages/tensorflow/contrib/framework/python/ops/arg_scope.py", line 181, in func_with_args\n    return func(*args, **current_args)', 'File "/media/bruce/MoreData/malcolm/tpc/base-pixel-cnn/pixel_cnn_pp/nn.py", line 238, in conv2d\n    tf.assert_variables_initialized([V, g, b])', 'File "/home/malcolm/data/anaconda3/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 170, in wrapped\n    return _add_should_use_warning(fn(*args, **kwargs))', 'File "/home/malcolm/data/anaconda3/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 139, in _add_should_use_warning\n    wrapped = TFShouldUseWarningWrapper(x)', 'File "/home/malcolm/data/anaconda3/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 96, in __init__\n    stack = [s.strip() for s in traceback.format_stack()]']
==================================

It still seems to run OK after printing the errors, though. The culprit seems to be this line, which appears three times in pixel_cnn_pp/nn.py:

tf.assert_variables_initialized([V, g, b])

It seems that this function expects its return value to be used, and that TF 1.2 introduced a check to print an error if the return value isn't used. As a workaround, commenting out this line (in each place it appears) removes the errors.

Jun 27 '17 23:06 malcolmsharpe

I am experiencing this too.

Jul 04 '17 20:07 hanss314

me too.

Jul 07 '17 08:07 weixsong

me three, here is a bit more context on how I came across this error:

    input_ = tf.placeholder(tf.int32, [None, None], name='input')
    targets = tf.placeholder(tf.int32, [None, None], name='target')
    learning_rate = tf.placeholder(tf.float32, name='learning_rate')
    keep_prob = tf.placeholder(tf.float32, name='keep_prob')
    len_source = tf.placeholder(tf.int32, [None], name='source_sequence_length')
    len_target = tf.placeholder(tf.int32, [None], name='target_sequence_length')
    max_target = tf.reduce_max(len_target, name='max_target_len')
    
    return (input_, targets, learning_rate, keep_prob, len_target, max_target, len_source)

I ran the above in a jupyter cell and got the error.

Jul 08 '17 20:07 ksooklall

me four. Any official solutions?

Aug 03 '17 08:08 xyc1207

me five, any solution wor work arround?

Aug 15 '17 22:08 tlan16

I've got this error too. I think you all would better to check that the function which got issue whether deprecated or not. In my case, the function 'initialize_all_variables' was deprecated in version 1.3.

Aug 17 '17 08:08 JINwonLEE

I was able to solve the problem for my particular use case by passing a "shape" for scalar placeholders. So something like this:

learning_rate = tf.placeholder(tf.float32, name='learning_rate', shape=())
keep_prob = tf.placeholder(tf.float32, name='keep_prob', shape=())

instead of

learning_rate = tf.placeholder(tf.float32, name='learning_rate')
keep_prob = tf.placeholder(tf.float32, name='keep_prob')

I believe what was happening (in my case), was that I had unit tests making assertions against the shape of these variables but not using them, hence the error. There must have been an extra validation check added for this in tensorflow 1.2

Aug 17 '17 16:08 calvinalvin

I have same issues with this

my placeholder define like this:

` global_step = tf.get_variable('global_step', [],
                                      initializer = tf.constant_initializer(0),dtype=tf.int32)
        # global_step = tf.Variable(0,name='global_step')
        #global_step=tf.get_variable('global_step', shape=[], initializer=tf.zeros_initializer(), dtype=tf.int32, trainable=False)

        x = tf.placeholder(tf.float32,shape=[None,num_input],name="mfcc_input")
        y = tf.placeholder(tf.int32,shape=[None,num_classes],name="labels")
        is_training = tf.placeholder(dtype=bool, shape=[], name="is_training")
        q_selector = tf.cond(is_training,lambda:[train_mfcc_batch,train_labels_batch],lambda:[test_mfcc_batch,test_labels_batch])`

and it just shows this error when I try to run on the different machine with GPUs (distributed tensorflow):

`2017-08-25 12:02:50.917393: W tensorflow/core/framework/op_kernel.cc:1148] Invalid argument: Shape [-1,13] has negative dimensions
2017-08-25 12:02:50.917476: E tensorflow/core/common_runtime/executor.cc:644] Executor failed to create kernel. Invalid argument: Shape [-1,13] has negative dimensions
    [[Node: mfcc_input = Placeholder[dtype=DT_FLOAT, shape=[?,13], _device="/job:worker/replica:0/task:0/gpu:0"]()]]
2017-08-25 12:02:50.918477: W tensorflow/core/framework/op_kernel.cc:1148] Invalid argument: Shape [-1,13] has negative dimensions
2017-08-25 12:02:50.918525: E tensorflow/core/common_runtime/executor.cc:644] Executor failed to create kernel. Invalid argument: Shape [-1,13] has negative dimensions
   [[Node: mfcc_input = Placeholder[dtype=DT_FLOAT, shape=[?,13], _device="/job:worker/replica:0/task:0/gpu:0"]()]]
2017-08-25 12:02:50.920048: W tensorflow/core/framework/op_kernel.cc:1148] Invalid argument: Shape [-1,3] has negative dimensions
2017-08-25 12:02:50.920092: E tensorflow/core/common_runtime/executor.cc:644] Executor failed to create kernel. Invalid argument: Shape [-1,3] has negative dimensions
   [[Node: labels = Placeholder[dtype=DT_INT32, shape=[?,3], _device="/job:worker/replica:0/task:0/gpu:0"]()]]
ERROR:tensorflow:==================================
Object was never used (type ):

If you want to mark it as used call its "mark_used()" method.
It was originally created here:
['File "sound_classifier_distributed_supervisor.py", line 164, in \n    train_op = rep_op.minimize(xent,global_step=global_step)', 'File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/optimizer.py", line 325, in minimize\n    name=name)', 'File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/sync_replicas_optimizer.py", line 252, in apply_gradients\n    variables.global_variables())', 'File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/tf_should_use.py", line 170, in wrapped\n    return _add_should_use_warning(fn(*args, **kwargs))', 'File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/tf_should_use.py", line 139, in _add_should_use_warning\n    wrapped = TFShouldUseWarningWrapper(x)', 'File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/tf_should_use.py", line 96, in __init__\n    stack = [s.strip() for s in traceback.format_stack()]']`