I am following along with a tutorial, and I am building a simple regression model with tensorflow. I would expect tf to fit the model without any hiccups. Instead, I am getting a value error.
The model and compile steps look identical to the tutorial.
The data is similar (two numpy arrays). I used different numbers in the arrays, but I do not think that is the issue. Any two arrays of equal length should be valid, right?
X = tf.constant(np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]))
y = tf.constant(np.array([1, 4, 7, 10, 13, 16, 19, 22, 25, 28, 31]))
model = tf.keras.Sequential([
tf.keras.layers.Dense(1)
])
model.compile(
loss=tf.keras.losses.mae,
optimizer=tf.keras.optimizers.SGD(),
metrics=["mae"]
)
model.fit(X, y, epochs=10)
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-41-2da4b2bd3c5c> in <module>
12 )
13
---> 14 model.fit(X, y, epochs=10)
~/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py in error_handler(*args, **kwargs)
59 def error_handler(*args, **kwargs):
60 if not tf.debugging.is_traceback_filtering_enabled():
---> 61 return fn(*args, **kwargs)
62
63 filtered_tb = None
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py in fit(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)
1562 ):
1563 callbacks.on_train_batch_begin(step)
-> 1564 tmp_logs = self.train_function(iterator)
1565 if data_handler.should_sync:
1566 context.async_wait()
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/util/traceback_utils.py in error_handler(*args, **kwargs)
139 try:
140 if not is_traceback_filtering_enabled():
--> 141 return fn(*args, **kwargs)
142 except NameError:
143 # In some very rare cases,
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/eager/def_function.py in __call__(self, *args, **kwds)
913
914 with OptionalXlaContext(self._jit_compile):
--> 915 result = self._call(*args, **kwds)
916
917 new_tracing_count = self.experimental_get_tracing_count()
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/eager/def_function.py in _call(self, *args, **kwds)
961 # This is the first call of __call__, so we have to initialize.
962 initializers = []
--> 963 self._initialize(args, kwds, add_initializers_to=initializers)
964 finally:
965 # At this point we know that the initialization is complete (or less
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/eager/def_function.py in _initialize(self, args, kwds, add_initializers_to)
783 self._graph_deleter = FunctionDeleter(self._lifted_initializer_graph)
784 self._concrete_stateful_fn = (
--> 785 self._stateful_fn._get_concrete_function_internal_garbage_collected( # pylint: disable=protected-access
786 *args, **kwds))
787
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/eager/function.py in _get_concrete_function_internal_garbage_collected(self, *args, **kwargs)
2521 args, kwargs = None, None
2522 with self._lock:
-> 2523 graph_function, _ = self._maybe_define_function(args, kwargs)
2524 return graph_function
2525
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/eager/function.py in _maybe_define_function(self, args, kwargs)
2758 # Only get placeholders for arguments, not captures
2759 args, kwargs = placeholder_dict["args"]
-> 2760 graph_function = self._create_graph_function(args, kwargs)
2761
2762 graph_capture_container = graph_function.graph._capture_func_lib # pylint: disable=protected-access
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/eager/function.py in _create_graph_function(self, args, kwargs)
2668 arg_names = base_arg_names missing_arg_names
2669 graph_function = ConcreteFunction(
-> 2670 func_graph_module.func_graph_from_py_func(
2671 self._name,
2672 self._python_function,
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/framework/func_graph.py in func_graph_from_py_func(name, python_func, args, kwargs, signature, func_graph, autograph, autograph_options, add_control_dependencies, arg_names, op_return_value, collections, capture_by_value, acd_record_initial_resource_uses)
1245 _, original_func = tf_decorator.unwrap(python_func)
1246
-> 1247 func_outputs = python_func(*func_args, **func_kwargs)
1248
1249 # invariant: `func_outputs` contains only Tensors, CompositeTensors,
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/eager/def_function.py in wrapped_fn(*args, **kwds)
675 # the function a weak reference to itself to avoid a reference cycle.
676 with OptionalXlaContext(compile_with_xla):
--> 677 out = weak_wrapped_fn().__wrapped__(*args, **kwds)
678 return out
679
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/framework/func_graph.py in autograph_handler(*args, **kwargs)
1231 except Exception as e: # pylint:disable=broad-except
1232 if hasattr(e, "ag_error_metadata"):
-> 1233 raise e.ag_error_metadata.to_exception(e)
1234 else:
1235 raise
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/framework/func_graph.py in autograph_handler(*args, **kwargs)
1220 # TODO(mdan): Push this block higher in tf.function's call stack.
1221 try:
-> 1222 return autograph.converted_call(
1223 original_func,
1224 args,
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py in converted_call(f, args, kwargs, caller_fn_scope, options)
437 try:
438 if kwargs is not None:
--> 439 result = converted_f(*effective_args, **kwargs)
440 else:
441 result = converted_f(*effective_args)
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py in tf__train_function(iterator)
13 try:
14 do_return = True
---> 15 retval_ = ag__.converted_call(ag__.ld(step_function), (ag__.ld(self), ag__.ld(iterator)), None, fscope)
16 except:
17 do_return = False
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py in converted_call(f, args, kwargs, caller_fn_scope, options)
375
376 if not options.user_requested and conversion.is_allowlisted(f):
--> 377 return _call_unconverted(f, args, kwargs, options)
378
379 # internal_convert_user_code is for example turned off when issuing a dynamic
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py in _call_unconverted(f, args, kwargs, options, update_cache)
457 if kwargs is not None:
458 return f(*args, **kwargs)
--> 459 return f(*args)
460
461
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py in step_function(model, iterator)
1144 )
1145 data = next(iterator)
-> 1146 outputs = model.distribute_strategy.run(run_step, args=(data,))
1147 outputs = reduce_per_replica(
1148 outputs, self.distribute_strategy, reduction="first"
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py in run(***failed resolving arguments***)
1313 fn = autograph.tf_convert(
1314 fn, autograph_ctx.control_status_ctx(), convert_by_default=False)
-> 1315 return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
1316
1317 def reduce(self, reduce_op, value, axis):
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py in call_for_each_replica(self, fn, args, kwargs)
2889 kwargs = {}
2890 with self._container_strategy().scope():
-> 2891 return self._call_for_each_replica(fn, args, kwargs)
2892
2893 def _call_for_each_replica(self, fn, args, kwargs):
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py in _call_for_each_replica(self, fn, args, kwargs)
3690 def _call_for_each_replica(self, fn, args, kwargs):
3691 with ReplicaContext(self._container_strategy(), replica_id_in_sync_group=0):
-> 3692 return fn(*args, **kwargs)
3693
3694 def _reduce_to(self, reduce_op, value, destinations, options):
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py in wrapper(*args, **kwargs)
687 try:
688 with conversion_ctx:
--> 689 return converted_call(f, args, kwargs, options=options)
690 except Exception as e: # pylint:disable=broad-except
691 if hasattr(e, 'ag_error_metadata'):
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py in converted_call(f, args, kwargs, caller_fn_scope, options)
375
376 if not options.user_requested and conversion.is_allowlisted(f):
--> 377 return _call_unconverted(f, args, kwargs, options)
378
379 # internal_convert_user_code is for example turned off when issuing a dynamic
~/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py in _call_unconverted(f, args, kwargs, options, update_cache)
456
457 if kwargs is not None:
--> 458 return f(*args, **kwargs)
459 return f(*args)
460
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py in run_step(data)
1133
1134 def run_step(data):
-> 1135 outputs = model.train_step(data)
1136 # Ensure counter is updated only if `train_step` succeeds.
1137 with tf.control_dependencies(_minimum_control_deps(outputs)):
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py in train_step(self, data)
991 # Run forward pass.
992 with tf.GradientTape() as tape:
--> 993 y_pred = self(x, training=True)
994 loss = self.compute_loss(x, y, y_pred, sample_weight)
995 self._validate_target_and_loss(y, loss)
~/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py in error_handler(*args, **kwargs)
59 def error_handler(*args, **kwargs):
60 if not tf.debugging.is_traceback_filtering_enabled():
---> 61 return fn(*args, **kwargs)
62
63 filtered_tb = None
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py in __call__(self, *args, **kwargs)
555 layout_map_lib._map_subclass_model_variable(self, self._layout_map)
556
--> 557 return super().__call__(*args, **kwargs)
558
559 @doc_controls.doc_in_current_and_subclasses
~/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py in error_handler(*args, **kwargs)
59 def error_handler(*args, **kwargs):
60 if not tf.debugging.is_traceback_filtering_enabled():
---> 61 return fn(*args, **kwargs)
62
63 filtered_tb = None
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/base_layer.py in __call__(self, *args, **kwargs)
1095 self._compute_dtype_object
1096 ):
-> 1097 outputs = call_fn(inputs, *args, **kwargs)
1098
1099 if self._activity_regularizer:
~/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py in error_handler(*args, **kwargs)
153 else:
154 new_e = e
--> 155 raise new_e.with_traceback(e.__traceback__) from None
156 finally:
157 del signature
~/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py in error_handler(*args, **kwargs)
94 bound_signature = None
95 try:
---> 96 return fn(*args, **kwargs)
97 except Exception as e:
98 if hasattr(e, "_keras_call_info_injected"):
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/sequential.py in call(self, inputs, training, mask)
423 kwargs["training"] = training
424
--> 425 outputs = layer(inputs, **kwargs)
426
427 if len(tf.nest.flatten(outputs)) != 1:
~/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py in error_handler(*args, **kwargs)
59 def error_handler(*args, **kwargs):
60 if not tf.debugging.is_traceback_filtering_enabled():
---> 61 return fn(*args, **kwargs)
62
63 filtered_tb = None
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/base_layer.py in __call__(self, *args, **kwargs)
1063 ):
1064
-> 1065 input_spec.assert_input_compatibility(
1066 self.input_spec, inputs, self.name
1067 )
~/opt/anaconda3/lib/python3.8/site-packages/keras/engine/input_spec.py in assert_input_compatibility(input_spec, inputs, layer_name)
248 ndim = x.shape.rank
249 if ndim is not None and ndim < spec.min_ndim:
--> 250 raise ValueError(
251 f'Input {input_index} of layer "{layer_name}" '
252 "is incompatible with the layer: "
ValueError: in user code:
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py", line 1160, in train_function *
return step_function(self, iterator)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py", line 1146, in step_function **
outputs = model.distribute_strategy.run(run_step, args=(data,))
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py", line 1315, in run
return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py", line 2891, in call_for_each_replica
return self._call_for_each_replica(fn, args, kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/tensorflow/python/distribute/distribute_lib.py", line 3692, in _call_for_each_replica
return fn(*args, **kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py", line 1135, in run_step **
outputs = model.train_step(data)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py", line 993, in train_step
y_pred = self(x, training=True)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 61, in error_handler
return fn(*args, **kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/training.py", line 557, in __call__
return super().__call__(*args, **kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 61, in error_handler
return fn(*args, **kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/base_layer.py", line 1097, in __call__
outputs = call_fn(inputs, *args, **kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 155, in error_handler
raise new_e.with_traceback(e.__traceback__) from None
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 96, in error_handler
return fn(*args, **kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/sequential.py", line 425, in call
outputs = layer(inputs, **kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/utils/traceback_utils.py", line 61, in error_handler
return fn(*args, **kwargs)
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/base_layer.py", line 1065, in __call__
input_spec.assert_input_compatibility(
File "/Users/mcm66103/opt/anaconda3/lib/python3.8/site-packages/keras/engine/input_spec.py", line 250, in assert_input_compatibility
raise ValueError(
ValueError: Exception encountered when calling layer "sequential_24" " f"(type Sequential).
Input 0 of layer "dense_25" is incompatible with the layer: expected min_ndim=2, found ndim=1. Full shape received: (None,)
Call arguments received by layer "sequential_24" " f"(type Sequential):
• inputs=tf.Tensor(shape=(None,), dtype=int64)
• training=True
• mask=None
CodePudding user response：
You are missing the feature dimension necessary for the Dense layer, since your model is inferring the input shape based on the data that you feed, so try:
import tensorflow as tf
import numpy as np
X = tf.constant(np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]))[:, None]
y = tf.constant(np.array([1, 4, 7, 10, 13, 16, 19, 22, 25, 28, 31]))
model = tf.keras.Sequential([
tf.keras.layers.Dense(1)
])
model.compile(
loss=tf.keras.losses.mae,
optimizer=tf.keras.optimizers.SGD(),
metrics=["mae"]
)
model.fit(X, y, epochs=10)
You could also use X = tf.expand_dims(X, axis=-1).
CodePudding user response：
Try adding an input layer in the Sequential model:
X = tf.constant(np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]))
y = tf.constant(np.array([1, 4, 7, 10, 13, 16, 19, 22, 25, 28, 31]))
model = tf.keras.Sequential([
tf.keras.layers.InputLayer(input_shape=(1,)), # This will take care of dimensions
tf.keras.layers.Dense(1)
])
model.compile(
loss=tf.keras.losses.mae,
optimizer=tf.keras.optimizers.SGD(),
metrics=["mae"]
)
model.fit(X, y, epochs=10)
For reference see: tf.keras.InputLayer
CodePudding user response：
You should try this code...
X = tf.constant(np.array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]))
y = tf.constant(np.array([1, 4, 7, 10, 13, 16, 19, 22, 25, 28, 31]))
model = tf.keras.Sequential([
tf.keras.layers.Dense(1)
])
model.compile(
loss=tf.keras.losses.mae,
optimizer=tf.keras.optimizers.SGD(),
metrics=["mae"]
)
X = tf.expand_dims(X , axis=0)
y = tf.expand_dims(y , axis=0)
model.fit(X, y, epochs=10)
Output
Epoch 1/10
1/1 [==============================] - 1s 777ms/step - loss: 14.0288 - mae: 14.0288
Epoch 2/10
1/1 [==============================] - 0s 8ms/step - loss: 11.6894 - mae: 11.6894
Epoch 3/10
1/1 [==============================] - 0s 9ms/step - loss: 10.2720 - mae: 10.2720
Epoch 4/10
1/1 [==============================] - 0s 9ms/step - loss: 9.4745 - mae: 9.4745
Epoch 5/10
1/1 [==============================] - 0s 9ms/step - loss: 8.9153 - mae: 8.9153
Epoch 6/10
1/1 [==============================] - 0s 10ms/step - loss: 8.6282 - mae: 8.6282
Epoch 7/10
1/1 [==============================] - 0s 9ms/step - loss: 8.4167 - mae: 8.4167
Epoch 8/10
1/1 [==============================] - 0s 5ms/step - loss: 8.3848 - mae: 8.3848
Epoch 9/10
1/1 [==============================] - 0s 9ms/step - loss: 8.3529 - mae: 8.3529
Epoch 10/10
1/1 [==============================] - 0s 10ms/step - loss: 8.3210 - mae: 8.3210