[ Dy2Static ] Fix bugs when select inputs meeting different shape or...

[ Dy2Static ] Fix bugs when select inputs meeting different shape or undefined-var (#45916) (#46020)

* fix select_input with different shape errors:
1. select_input_with_buildin_type directly return non-undefinedvar branch when meeting undefined var
2. the output shape of select_input is inferred from inputs.

* reverse the logic in select_input

python/paddle/fluid/dygraph/dygraph_to_static/utils.py

@@ -145,21 +145,6 @@ def data_layer_not_check(name, shape, dtype='float32', lod_level=0):
                                          need_check_feed=False)
 
 
-def create_undefined_var_like(variable):
-    """ create a undefined var with the same shape and dtype like varaible.
-    """
-    from paddle.fluid.dygraph.dygraph_to_static.return_transformer import RETURN_NO_VALUE_MAGIC_NUM
-    var = data_layer_not_check(unique_name.generate("undefined_var"),
-                               variable.shape, variable.dtype)
-    var.stop_gradient = False
-    helper = LayerHelper('create_undefined_var_like', **locals())
-    saved_block_ids = helper.main_program.current_block_idx
-    helper.main_program.current_block_idx = 0
-    assign(RETURN_NO_VALUE_MAGIC_NUM, var)
-    helper.main_program.current_block_idx = saved_block_ids
-    return var
-
-
 def create_undefined_variable():
     from paddle.fluid.dygraph.dygraph_to_static.return_transformer import RETURN_NO_VALUE_MAGIC_NUM
     var = data_layer_not_check(unique_name.generate("undefined_var"), [1],

python/paddle/fluid/layers/control_flow.py

@@ -70,6 +70,25 @@ def select_output(input, outputs, mask):
     return outputs
 
 
+def _select_input_infer_shape(first_shape, second_shape):
+    """
+    This function infer the output shape by following algorithm:
+    1. if the dims is different, raise a error.
+    2. compare axis one by one:
+        if a == b: we set axis to a
+        if a != b: we set axis to -1
+    for compatibility，non declarative mode, we just return second_shape.
+    """
+    if len(first_shape) != len(second_shape):
+        warnings.warn(
+            f"the input shapes of select_input should have the same rank, but get {first_shape}, {second_shape}"
+        )
+        return second_shape
+    out_shape = list(
+        map(lambda a, b: a if a == b else -1, first_shape, second_shape))
+    return out_shape
+
+
 def select_input(inputs, mask):
     """
     **select_input**
@@ -89,13 +108,15 @@ def select_input(inputs, mask):
     check_type(inputs, 'inputs', (list, tuple), 'select_input')
     check_variable_and_dtype(mask, 'mask', ['int32'], 'select_input')
 
-    input_dtype = inputs[1].dtype
-    input_shape = inputs[1].shape
-    input_type = inputs[1].type
+    # Select input should expand the shape. If it is - 1 and valid number, use - 1 first. If the dim is different, an error will be reported directly
+    #assert inputs[0].dtype == inputs[1].dtype, f"Expect the inputs should have the same dtype, but get {inputs[0].dtype} and {inputs[1].dtype}"
+    output_shape = _select_input_infer_shape(inputs[0].shape, inputs[1].shape)
+    output_dtype = inputs[1].dtype
+    output_type = inputs[1].type
 
-    out = helper.create_variable(dtype=input_dtype,
-                                 shape=input_shape,
-                                 type=input_type)
+    out = helper.create_variable(dtype=output_dtype,
+                                 shape=output_shape,
+                                 type=output_type)
     helper.append_op(type='select_input',
                      inputs={
                          'X': inputs,
@@ -105,9 +126,9 @@ def select_input(inputs, mask):
     return out
 
 
-def select_input_with_buildin_type(inputs, mask):
+def select_input_with_buildin_type(inputs, mask, name):
     from paddle.fluid.dygraph.dygraph_to_static.variable_trans_func import to_static_variable
-    from paddle.fluid.dygraph.dygraph_to_static.utils import UndefinedVar, create_undefined_var_like
+    from paddle.fluid.dygraph.dygraph_to_static.utils import UndefinedVar
     false_var, true_var = inputs
 
     if isinstance(false_var, UndefinedVar) and isinstance(
@@ -117,7 +138,11 @@ def select_input_with_buildin_type(inputs, mask):
         return None
 
     if isinstance(false_var, Variable) and isinstance(true_var, Variable):
-        return select_input(inputs, mask)
+        try:
+            return select_input(inputs, mask)
+        except Exception as e:
+            raise RuntimeError(
+                f"Exceptions throwed while doing select_input on {name}:\n{e}")
 
     elif (isinstance(false_var, (support_ret_buildin_type))
           and isinstance(false_var, type(true_var))):
@@ -148,24 +173,19 @@ def select_input_with_buildin_type(inputs, mask):
             if isinstance(a, UndefinedVar): return a
             return to_static_variable(a)
 
-        def create_like_if_undefined_var(a, b):
-            if isinstance(a, UndefinedVar): return create_undefined_var_like(b)
-            return a
-
-        # TODO(xiongkun): add warning here.
-        true_var, false_var = create_var_if_not_undefined_var(
-            true_var), create_var_if_not_undefined_var(false_var)
-        inputs = [
-            create_like_if_undefined_var(false_var, true_var),
-            create_like_if_undefined_var(true_var, false_var)
-        ]
+        true_var, false_var = to_static_variable(true_var), to_static_variable(
+            false_var)
+        inputs = [false_var, true_var]
     else:
         raise TypeError(
             "Unsupported return type of true_fn and false_fn in cond: false_var "
             "returned by fasle_fn is '{}' and true_var of true_fn is '{}'".
             format(type(false_var), type(true_var)))
-
-    return select_input(inputs, mask)
+    try:
+        return select_input(inputs, mask)
+    except Exception as e:
+        raise RuntimeError(
+            f"Exceptions throwed while doing select_input on {name}:\n{e}")
 
 
 def split_lod_tensor(input, mask, level=0):
@@ -2658,9 +2678,16 @@ def cond(pred, true_fn=None, false_fn=None, name=None, return_names=None):
                 .format(return_name, e))
 
     mask = cast(pred, dtype='int32')
-    merge_func = lambda false_var, true_var: select_input_with_buildin_type(
-        [false_var, true_var], mask)
-    merged_output = map_structure(merge_func, false_output, true_output)
+    merge_func = lambda name, false_var, true_var: select_input_with_buildin_type(
+        [false_var, true_var], mask, name)
+
+    def merge_every_var_list(false_vars, true_vars, name):
+        return map_structure(partial(merge_func, name), false_vars, true_vars)
+
+    merged_output = list(
+        map(merge_every_var_list, to_sequence(false_output),
+            to_sequence(true_output), to_sequence(return_names)))
+    merged_output = pack_sequence_as(false_output, flatten(merged_output))
     return merged_output
 
 

python/paddle/fluid/tests/unittests/dygraph_to_static/ifelse_simple_func.py

@@ -164,7 +164,7 @@ def nested_if_else(x_v):
             if paddle.mean(y).numpy()[0] < batch_size:
                 y = fluid.layers.abs(y)
             else:
-                tmp = fluid.layers.fill_constant([feat_size],
+                tmp = fluid.layers.fill_constant(y.shape,
                                                  dtype='float32',
                                                  value=-1)
                 y = y - tmp
@@ -273,7 +273,7 @@ class NetWithControlFlowIf(fluid.dygraph.Layer):
                         [hidden_dim], dtype='float32', value=9)
                     y = fluid.layers.abs(y)
                 else:
-                    tmp = fluid.layers.fill_constant([5],
+                    tmp = fluid.layers.fill_constant(y.shape,
                                                      dtype='float32',
                                                      value=-1)
                     y = y - tmp