Prediction Issue

[SilenceBSun]

My code:
grad.beta<-predict(model.grad, keras_train_data_x, keras_train_data_y)
grad.beta <- data.frame(grad.beta)

Hi, I am new user of Keras. When I try to predict the data with keras, it shows me this.

model is fine. model.grad has processed without any errors showing.
my tensorflow version is 2.1. numpy version is 1.16.4. python version is 3.6. keras version is 2.3.1.

[t-kalinowski]

• Can you please provide code I can run to reproduce the error locally? {reprex} may be helpful.
• What is the output from running reticulate::py_last_error()?
• Your tensorflow version and python versions are quite old, and no longer supported. You may want to upgrade (though, I don't think that's the source of the error).

[SilenceBSun]

── Python Exception Message ──────────────────────────────────────────────────────────
Traceback (most recent call last):
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\eager\execute.py", line 61, in quick_execute
    num_outputs)
TypeError: An op outside of the function building code is being passed
a "Graph" tensor. It is possible to have Graph tensors
leak out of the function building context by including a
tf.init_scope in your function building code.
For example, the following function will fail:
  @tf.function
  def has_init_scope():
    my_constant = tf.constant(1.)
    with tf.init_scope():
      added = my_constant * 2
The graph tensor has name: lambda_21/Identity:0

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\keras\engine\training.py", line 1013, in predict
    use_multiprocessing=use_multiprocessing)
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\keras\engine\training_v2.py", line 498, in predict
    workers=workers, use_multiprocessing=use_multiprocessing, **kwargs)
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\keras\engine\training_v2.py", line 475, in _model_iteration
    total_epochs=1)
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\keras\engine\training_v2.py", line 128, in run_one_epoch
    batch_outs = execution_function(iterator)
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\keras\engine\training_v2_utils.py", line 98, in execution_function
    distributed_function(input_fn))
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\eager\def_function.py", line 568, in __call__
    result = self._call(*args, **kwds)
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\eager\def_function.py", line 638, in _call
    return self._concrete_stateful_fn._filtered_call(canon_args, canon_kwds)  # pylint: disable=protected-access
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\eager\function.py", line 1611, in _filtered_call
    self.captured_inputs)
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\eager\function.py", line 1692, in _call_flat
    ctx, args, cancellation_manager=cancellation_manager))
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\eager\function.py", line 545, in call
    ctx=ctx)
  File "C:\Users\huany\MINICO~1\envs\TENSOR~1\lib\site-packages\tensorflow_core\python\eager\execute.py", line 75, in quick_execute
    "tensors, but found {}".format(keras_symbolic_tensors))
tensorflow.python.eager.core._SymbolicException: Inputs to eager execution function cannot be Keras symbolic tensors, but found [<tf.Tensor 'lambda_21/Identity:0' shape=(None,) dtype=float32>, <tf.Tensor 'Predictions_10/Identity:0' shape=(None, 3) dtype=float32>, <tf.Tensor 'lambda_21/strided_slice/stack:0' shape=(2,) dtype=int32>, <tf.Tensor 'lambda_21/strided_slice/stack_1:0' shape=(2,) dtype=int32>, <tf.Tensor 'lambda_21/strided_slice/stack_2:0' shape=(2,) dtype=int32>, <tf.Tensor 'Predictions_10/MatMul/ReadVariableOp:0' shape=(6, 3) dtype=float32>, <tf.Tensor 'dense_21/Identity:0' shape=(None, 6) dtype=float32>, <tf.Tensor 'dense_21/Tanh:0' shape=(None, 6) dtype=float32>, <tf.Tensor 'dense_21/MatMul/ReadVariableOp:0' shape=(12, 6) dtype=float32>, <tf.Tensor 'dense_22/Identity:0' shape=(None, 12) dtype=float32>, <tf.Tensor 'dense_22/Tanh:0' shape=(None, 12) dtype=float32>, <tf.Tensor 'dense_22/MatMul/ReadVariableOp:0' shape=(3, 12) dtype=float32>, <tf.Tensor 'input_11:0' shape=(None, 3) dtype=float32>]

── R Traceback ───────────────────────────────────────────────────────────────────────
    ▆
 1. ├─model.grad %>% predict(keras_train_data_x)
 2. ├─stats::predict(., keras_train_data_x)
 3. └─keras:::predict.keras.engine.training.Model(., keras_train_data_x)
 4.   ├─base::do.call(object$predict, args)
 5.   └─reticulate (local) `<python.builtin.method>`(...)
 6.     └─reticulate:::py_call_impl(callable, call_args$unnamed, call_args$named)

[SilenceBSun]

This is the result from the reticulate :: py_last_error(), actually, I have just changed some labels. It seems that I fix the last problem, however, this is a new problem occurred now. I cannot ready the reticulate::py_last_error() currently, I am going to try.

[SilenceBSun]

Actually, I find out it is a matrix type problem. I turn the data into matrix type (Original one is data frame). If I do not turn it into the a matrix type, the error will be the same as the first one.

[t-kalinowski]

If you can condense the issue into a block of code that I can run to reproduce the error, I'm happy to take a look. (e.g., generating fake data with runif() or similar in the reprex).

[SilenceBSun]

Hi, I have uploaded my code and data, which is made by myself. Because it is for my thesis, I prefer to delete it after you took a look. https://github.com/Vingoal/solid-octo-goggles

[t-kalinowski]

• the link doesn't work for me (I get 404)
• Surely you can condense the issue into an MRE (and, odds are good that while condensing it down to a minimal reproducible example, you resolve the issue yourself :)

[SilenceBSun]

Thank you for the reply, t-kalinowski. I am currently busy to work on other articles. Please give me some time to condense the issue into an MRE. I think I still need your help in the future. Have a nice day.