Issue in executing m.predict()

[ganeshjayaraman]

I receive the following error while executing m.predict()

The code associated with m.predict(input_fn=pred_fn) :

def pred_fn():
sample = [ 1, 127, 1, 3, 1683, 19, 26, 17, 475, 0, 9, 0, 3, "05db9164", "8947f767", "11c9d79e", "52a787c8", "4cf72387", "fbad5c96", "18671b18", "0b153874", "a73ee510", "ceb10289", "77212bd7", "79507c6b", "7203f04e", "07d13a8f", "2c14c412", "49013ffe", "8efede7f", "bd17c3da", "f6a3e43b", "a458ea53", "35cd95c9", "ad3062eb", "c7dc6720", "3fdb382b", "010f6491", "49d68486"]
sample_dict = dict(zip(FEATURE_COLUMNS, sample))

for feature_name in CATEGORICAL_COLUMNS:
    sample_dict[feature_name] = tf.expand_dims(sample_dict[feature_name], -1)
    
for feature_name in CONTINUOUS_COLUMNS:
    **sample_dict[feature_name] = tf.constant(sample_dict[feature_name], dtype=tf.int32)**  # I think here is the problem .. But i don't know how to resolve it ..
print(sample_dict)

return sample_dict

Using TensorFlow version 1.4.0
TensorFlow version 1.4.0
Begin training and evaluation
Saving model checkpoints to models/model_WIDE_AND_DEEP_1510910853
estimator built
2017-11-17 14:57:41.105234: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX
fit done
evaluate done
Accuracy: 0.7945

{'I12': <tf.Tensor 'Const_11:0' shape=() dtype=int32>, 'C12': <tf.Tensor 'ExpandDims_11:0' shape=(1,) dtype=string>, 'I4': <tf.Tensor 'Const_3:0' shape=() dtype=int32>, 'I9': <tf.Tensor 'Const_8:0' shape=() dtype=int32>, 'C20': <tf.Tensor 'ExpandDims_19:0' shape=(1,) dtype=string>, 'C25': <tf.Tensor 'ExpandDims_24:0' shape=(1,) dtype=string>, 'C6': <tf.Tensor 'ExpandDims_5:0' shape=(1,) dtype=string>, 'C17': <tf.Tensor 'ExpandDims_16:0' shape=(1,) dtype=string>, 'C4': <tf.Tensor 'ExpandDims_3:0' shape=(1,) dtype=string>, 'I1': <tf.Tensor 'Const:0' shape=() dtype=int32>, 'I8': <tf.Tensor 'Const_7:0' shape=() dtype=int32>, 'I3': <tf.Tensor 'Const_2:0' shape=() dtype=int32>, 'C26': <tf.Tensor 'ExpandDims_25:0' shape=(1,) dtype=string>, 'C5': <tf.Tensor 'ExpandDims_4:0' shape=(1,) dtype=string>, 'C24': <tf.Tensor 'ExpandDims_23:0' shape=(1,) dtype=string>, 'I7': <tf.Tensor 'Const_6:0' shape=() dtype=int32>, 'C16': <tf.Tensor 'ExpandDims_15:0' shape=(1,) dtype=string>, 'C19': <tf.Tensor 'ExpandDims_18:0' shape=(1,) dtype=string>, 'C18': <tf.Tensor 'ExpandDims_17:0' shape=(1,) dtype=string>, 'I11': <tf.Tensor 'Const_10:0' shape=() dtype=int32>, 'C23': <tf.Tensor 'ExpandDims_22:0' shape=(1,) dtype=string>, 'I5': <tf.Tensor 'Const_4:0' shape=() dtype=int32>, 'C14': <tf.Tensor 'ExpandDims_13:0' shape=(1,) dtype=string>, 'I6': <tf.Tensor 'Const_5:0' shape=() dtype=int32>, 'C10': <tf.Tensor 'ExpandDims_9:0' shape=(1,) dtype=string>, 'C21': <tf.Tensor 'ExpandDims_20:0' shape=(1,) dtype=string>, 'C3': <tf.Tensor 'ExpandDims_2:0' shape=(1,) dtype=string>, 'C7': <tf.Tensor 'ExpandDims_6:0' shape=(1,) dtype=string>, 'C15': <tf.Tensor 'ExpandDims_14:0' shape=(1,) dtype=string>, 'C2': <tf.Tensor 'ExpandDims_1:0' shape=(1,) dtype=string>, 'I2': <tf.Tensor 'Const_1:0' shape=() dtype=int32>, 'C11': <tf.Tensor 'ExpandDims_10:0' shape=(1,) dtype=string>, 'C8': <tf.Tensor 'ExpandDims_7:0' shape=(1,) dtype=string>, 'C9': <tf.Tensor 'ExpandDims_8:0' shape=(1,) dtype=string>, 'C13': <tf.Tensor 'ExpandDims_12:0' shape=(1,) dtype=string>, 'C1': <tf.Tensor 'ExpandDims:0' shape=(1,) dtype=string>, 'I13': <tf.Tensor 'Const_12:0' shape=() dtype=int32>, 'I10': <tf.Tensor 'Const_9:0' shape=() dtype=int32>, 'C22': <tf.Tensor 'ExpandDims_21:0' shape=(1,) dtype=string>}
Traceback (most recent call last):
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/layers/python/layers/feature_column_ops.py", line 127, in _input_from_feature_columns
transformed_tensor)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/layers/python/layers/feature_column.py", line 256, in _deep_embedding_lookup_arguments
"No deep embedding lookup arguments for column {}.".format(self))
NotImplementedError: No deep embedding lookup arguments for column _RealValuedColumn(column_name='I1', dimension=1, default_value=None, dtype=tf.float32, normalizer=None).

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/layers/python/layers/feature_column_ops.py", line 143, in _input_from_feature_columns
output_rank=output_rank))
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/layers/python/layers/feature_column.py", line 1828, in _to_dnn_input_layer
return _reshape_real_valued_tensor(input_tensor, output_rank, self.name)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/layers/python/layers/feature_column.py", line 1606, in _reshape_real_valued_tensor
raise ValueError(error_string)
ValueError: Error while processing column I1.Rank of input Tensor (0) should be the same as output_rank (2). For example, sequence data should typically be 3 dimensional (rank 3) while non-sequence data is typically 2 dimensional (rank 2).

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "task.py", line 320, in
train_and_eval(args.job_dir)
File "task.py", line 263, in train_and_eval
ans = m.predict(input_fn=pred_fn)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/util/deprecation.py", line 381, in new_func
return func(*args, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/util/deprecation.py", line 381, in new_func
return func(*args, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py", line 736, in predict
as_iterable=as_iterable)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/util/deprecation.py", line 381, in new_func
return func(*args, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py", line 771, in predict_classes
as_iterable=as_iterable)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/util/deprecation.py", line 316, in new_func
return func(*args, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 615, in predict
as_iterable=as_iterable)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 913, in _infer_model
infer_ops = self._get_predict_ops(features)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 1252, in _get_predict_ops
return self._call_model_fn(features, labels, model_fn_lib.ModeKeys.INFER)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 1166, in _call_model_fn
model_fn_results = self._model_fn(features, labels, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py", line 242, in _dnn_linear_combined_model_fn
scope=dnn_input_scope)
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/layers/python/layers/feature_column_ops.py", line 212, in input_from_feature_columns
default_name='input_from_feature_columns')
File "/usr/local/lib/python3.5/dist-packages/tensorflow/contrib/layers/python/layers/feature_column_ops.py", line 146, in _input_from_feature_columns
'{}, {}'.format(column.name, e, ee))
ValueError: Error creating input layer for column: I1.
Error while processing column I1.Rank of input Tensor (0) should be the same as output_rank (2). For example, sequence data should typically be 3 dimensional (rank 3) while non-sequence data is typically 2 dimensional (rank 2)., No deep embedding lookup arguments for column _RealValuedColumn(column_name='I1', dimension=1, default_value=None, dtype=tf.float32, normalizer=None).

From the error i can understand one thing the input rank is scalar but it should be a matrix .. Correct me if i'm wrong and i'm a newbie here in tensorflow

[githubier]

ok, I find change the
sample_dict[feature_name] = tf.constant(sample_dict[feature_name], dtype=tf.int32)
to
sample_dict[feature_name] = tf.expand_dims(tf.constant(sample_dict[feature_name], dtype=tf.int32), -1)