Remove Input requirement in dygraph for Model (#27557)

* remove input requirment in dygraph Model

* correct unittest

* upadte save inference model in dygraph without input

* fix unittets for test_model.py

* solve conflicts

* solve conflicts

* delete http.log

* fix test_model.py bug, correct initialization of MyModel

* fix unittests bugs

* set paddle manual seed for unittest

* fix Model bugs, because inputs can be list or dict when it is provided.

* add random seed for test_export_deploy_model

* delete redundant codes, because  calls

* Code optimization, error information optimization

python/paddle/hapi/model.py

@@ -200,6 +200,15 @@ def prepare_distributed_context(place=None):
     return strategy
 
 
+def _update_input_shapes(inputs):
+    shapes = None
+    if isinstance(inputs, list):
+        shapes = [list(input.shape) for input in inputs]
+    elif isinstance(inputs, dict):
+        shapes = [list(inputs[name].shape) for name in inputs]
+    return shapes
+
+
 class StaticGraphAdapter(object):
     """
     Model traning/inference with a static graph.
@@ -598,6 +607,7 @@ class DynamicGraphAdapter(object):
             'test_batch': 0
         }
 
+        self._input_shapes = None
         if self._nranks > 1:
             stradegy = fluid.dygraph.parallel.ParallelStrategy()
             stradegy.nranks = ParallelEnv().nranks
@@ -622,6 +632,7 @@ class DynamicGraphAdapter(object):
         self.model.network.train()
         self.mode = 'train'
         inputs = to_list(inputs)
+        self._input_shapes = _update_input_shapes(inputs)
         labels = labels or []
         labels = [to_variable(l) for l in to_list(labels)]
 
@@ -656,6 +667,7 @@ class DynamicGraphAdapter(object):
         self.model.network.eval()
         self.mode = 'eval'
         inputs = to_list(inputs)
+        self._input_shapes = _update_input_shapes(inputs)
         labels = labels or []
         labels = [to_variable(l) for l in to_list(labels)]
 
@@ -704,6 +716,7 @@ class DynamicGraphAdapter(object):
         self.model.network.eval()
         self.mode = 'test'
         inputs = [to_variable(x) for x in to_list(inputs)]
+        self._input_shapes = _update_input_shapes(inputs)
         outputs = self.model.network.forward(*inputs)
         if self._nranks > 1 and isinstance(self.model._place, fluid.CUDAPlace):
             outputs = [_all_gather(o, self._nranks) for o in to_list(outputs)]
@@ -778,7 +791,7 @@ class DynamicGraphAdapter(object):
 
         if not hasattr(self.model._optimizer, 'set_state_dict'):
             warnings.warn(
-                "paddle.fluid.optimizer is deprecated in API 2.0, please use paddle.optimizer instead"
+                "paddle.fluid.optimizer is deprecated in API 2.0, please use paddle.optimizer instead."
             )
             self.model._optimizer.set_dict(converted_state)
         else:
@@ -792,14 +805,15 @@ class Model(object):
     switched by `paddle.disable_static()`. The usage is as follows.
     But note, the switching between dynamic and static should be before
     instantiating a Model. The input description, i.e, paddle.static.InputSpec,
-    must be required.
+    must be required for static graph.
 
     Args:
         network (paddle.nn.Layer): The network is an instance of
             paddle.nn.Layer.
         inputs (InputSpec|list|dict|None): `inputs`, entry points of network,
             could be a InputSpec instance, or lits of InputSpec instances,
-            or dict ({name: InputSpec}), and it couldn't be None.
+            or dict ({name: InputSpec}), and it couldn't be None in static
+            graph.
         labels (InputSpec|list|None): `labels`, entry points of network,
             could be a InputSpec instnace or lits of InputSpec instances,
             or None. For static graph, if labels is required in loss,
@@ -844,14 +858,18 @@ class Model(object):
         self._loss = None
         self._loss_weights = None
         self._optimizer = None
-        self._optimizer = None
+        self._input_shapes = None
+        self._is_shape_inferred = False
         self._test_dataloader = None
 
+        if not in_dygraph_mode():
             if not isinstance(inputs, (list, dict, Input)):
                 raise TypeError(
-                "'inputs' must be list or dict in static graph mode")
+                    "'inputs' must be list or dict, and couldn't be None.")
+        elif inputs:
+            self._input_shapes = _update_input_shapes(inputs)
 
-        self._inputs = self._verify_spec(inputs, True)
+        self._inputs = self._verify_spec(inputs, is_input=True)
         self._labels = self._verify_spec(labels)
 
         # init backend
@@ -902,7 +920,12 @@ class Model(object):
               loss = model.train_batch([data], [label])
               print(loss)
         """
-        return self._adapter.train_batch(inputs, labels)
+        loss = self._adapter.train_batch(inputs, labels)
+        if fluid.in_dygraph_mode() and self._input_shapes is None:
+            self._input_shapes = self._adapter._input_shapes
+            self._is_shape_inferred = True
+            self._inputs = self._verify_spec(None, self._input_shapes, True)
+        return loss
 
     def eval_batch(self, inputs, labels=None):
         """
@@ -947,7 +970,12 @@ class Model(object):
               loss = model.eval_batch([data], [label])
               print(loss)
         """
-        return self._adapter.eval_batch(inputs, labels)
+        loss = self._adapter.eval_batch(inputs, labels)
+        if fluid.in_dygraph_mode() and self._input_shapes is None:
+            self._input_shapes = self._adapter._input_shapes
+            self._is_shape_inferred = True
+            self._inputs = self._verify_spec(None, self._input_shapes, True)
+        return loss
 
     def test_batch(self, inputs):
         """
@@ -987,7 +1015,12 @@ class Model(object):
               out = model.test_batch([data])
               print(out)
         """
-        return self._adapter.test_batch(inputs)
+        loss = self._adapter.test_batch(inputs)
+        if fluid.in_dygraph_mode() and self._input_shapes is None:
+            self._input_shapes = self._adapter._input_shapes
+            self._is_shape_inferred = True
+            self._inputs = self._verify_spec(None, self._input_shapes, True)
+        return loss
 
     def save(self, path, training=True):
         """  
@@ -1677,6 +1710,14 @@ class Model(object):
         if fluid.in_dygraph_mode():
             with fluid.framework._dygraph_guard(None):
                 layer = self.network
+                if self._input_shapes is None:  # No provided or inferred
+                    raise RuntimeError(
+                        "Saving inference model needs 'inputs' or running before saving. Please specify 'inputs' in Model initialization or input training zqqdata and perform a training for shape derivation."
+                    )
+                if self._is_shape_inferred:
+                    warnings.warn(
+                        "'inputs' was not specified when Model initialization, so the input shape to be saved will be the shape derived from the user's actual inputs. The input shape to be saved is %s. For saving correct input shapes, please provide 'inputs' for Model initialization."
+                        % self._input_shapes)
                 layer.forward = paddle.jit.to_static(
                     layer.forward, input_spec=self._inputs)
 
@@ -1775,6 +1816,7 @@ class Model(object):
             data = flatten(data)
             # LoDTensor.shape is callable, where LoDTensor comes from
             # DataLoader in static graph
+
             batch_size = data[0].shape()[0] if callable(data[
                 0].shape) else data[0].shape[0]
 
@@ -1864,10 +1906,26 @@ class Model(object):
             _input_size = self._inputs
         return summary(self.network, _input_size, dtype)
 
-    def _verify_spec(self, specs, is_input=False):
+    def _verify_spec(self, specs, shapes=None, is_input=False):
         out_specs = []
 
-        if isinstance(specs, dict):
+        if specs is None:
+            # Note(Aurelius84): If not specific specs of `Input`, using argument names of `forward` function
+            # to generate `Input`. But how can we know the actual shape of each input tensor?
+
+            if is_input:
+                arg_names = extract_args(self.network.forward)[1:]
+                if shapes is not None and fluid.in_dygraph_mode():
+                    out_specs = [
+                        Input(
+                            name=n, shape=shapes[i])
+                        for i, n in enumerate(arg_names)
+                    ]
+                else:
+                    out_specs = [Input(name=n, shape=[None]) for n in arg_names]
+            else:
+                out_specs = to_list(specs)
+        elif isinstance(specs, dict):
             assert is_input == False
             out_specs = [specs[n] \
                 for n in extract_args(self.network.forward) if n != 'self']

python/paddle/tests/test_model.py

@@ -66,34 +66,6 @@ class LeNetDygraph(paddle.nn.Layer):
         return x
 
 
-class LeNetDeclarative(fluid.dygraph.Layer):
-    def __init__(self, num_classes=10):
-        super(LeNetDeclarative, self).__init__()
-        self.num_classes = num_classes
-        self.features = Sequential(
-            Conv2d(
-                1, 6, 3, stride=1, padding=1),
-            ReLU(),
-            Pool2D(2, 'max', 2),
-            Conv2d(
-                6, 16, 5, stride=1, padding=0),
-            ReLU(),
-            Pool2D(2, 'max', 2))
-
-        if num_classes > 0:
-            self.fc = Sequential(
-                Linear(400, 120), Linear(120, 84), Linear(84, 10))
-
-    @declarative
-    def forward(self, inputs):
-        x = self.features(inputs)
-
-        if self.num_classes > 0:
-            x = fluid.layers.flatten(x, 1)
-            x = self.fc(x)
-        return x
-
-
 class MnistDataset(MNIST):
     def __init__(self, mode, return_label=True, sample_num=None):
         super(MnistDataset, self).__init__(mode=mode)
@@ -440,9 +412,7 @@ class TestModelFunction(unittest.TestCase):
         # dynamic saving
         device = paddle.set_device('cpu')
         fluid.enable_dygraph(device)
-        inputs = [InputSpec([None, 20], 'float32', 'x')]
-        labels = [InputSpec([None, 1], 'int64', 'label')]
-        model = Model(MyModel(), inputs, labels)
+        model = Model(MyModel())
         optim = fluid.optimizer.SGD(learning_rate=0.001,
                                     parameter_list=model.parameters())
         model.prepare(optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
@@ -545,6 +515,8 @@ class TestModelFunction(unittest.TestCase):
         paddle.summary(nlp_net, (1, 1, 2))
 
     def test_export_deploy_model(self):
+        self.set_seed()
+        np.random.seed(2020)
         for dynamic in [True, False]:
             paddle.disable_static() if dynamic else None
             prog_translator = ProgramTranslator()
@@ -579,6 +551,35 @@ class TestModelFunction(unittest.TestCase):
                 shutil.rmtree(save_dir)
             paddle.enable_static()
 
+    def test_dygraph_export_deploy_model_without_inputs(self):
+        mnist_data = MnistDataset(mode='train')
+        paddle.disable_static()
+        for initial in ["fit", "train_batch", "eval_batch", "test_batch"]:
+            save_dir = tempfile.mkdtemp()
+            if not os.path.exists(save_dir):
+                os.makedirs(save_dir)
+            net = LeNet()
+            model = Model(net)
+            optim = fluid.optimizer.Adam(
+                learning_rate=0.001, parameter_list=model.parameters())
+            model.prepare(
+                optimizer=optim, loss=CrossEntropyLoss(reduction="sum"))
+            if initial == "fit":
+                model.fit(mnist_data, batch_size=64, verbose=0)
+            else:
+                img = np.array(
+                    np.random.random((1, 1, 28, 28)), dtype=np.float32)
+                label = np.array(np.random.rand(1, 1), dtype=np.int64)
+                if initial == "train_batch":
+                    model.train_batch([img], [label])
+                elif initial == "eval_batch":
+                    model.eval_batch([img], [label])
+                else:
+                    model.test_batch([img])
+
+            model.save(save_dir, training=False)
+            shutil.rmtree(save_dir)
+
 
 class TestRaiseError(unittest.TestCase):
     def test_input_without_name(self):
@@ -589,12 +590,21 @@ class TestRaiseError(unittest.TestCase):
         with self.assertRaises(ValueError):
             model = Model(net, inputs, labels)
 
-    def test_input_without_input_spec(self):
-        for dynamic in [True, False]:
-            paddle.disable_static() if dynamic else None
+    def test_static_without_inputs(self):
+        paddle.enable_static()
         net = MyModel()
         with self.assertRaises(TypeError):
             model = Model(net)
+
+    def test_save_infer_model_without_inputs_and_run_in_dygraph(self):
+        paddle.disable_static()
+        net = MyModel()
+        save_dir = tempfile.mkdtemp()
+        if not os.path.exists(save_dir):
+            os.makedirs(save_dir)
+        with self.assertRaises(RuntimeError):
+            model = Model(net)
+            model.save(save_dir, training=False)
         paddle.enable_static()