[Dy2St] Make bool(Value) always throw error (PaddlePaddle#60902)

---------

Co-authored-by: zoooo0820 <zoooo0820@qq.com>

python/paddle/base/variable_index.py

@@ -82,12 +82,10 @@ def replace_none(item):
     return new_item, none_axes
 
 
-def is_integer_or_scalar_tensor(ele):
+def is_scalar_tensor(ele):
     from .framework import Variable
 
-    if type(ele) is int:
-        return True
-    elif isinstance(ele, Variable):
+    if isinstance(ele, Variable):
         if len(ele.shape) == 0 and ele.dtype != paddle.bool:
             return True
     elif isinstance(ele, paddle.pir.Value):
@@ -284,10 +282,9 @@ def parse_index(x, indices):
     dim = 0
     for i, slice_item in enumerate(indices):
         start, end, step = None, None, None
-        if is_integer_or_scalar_tensor(slice_item):
+        if type(slice_item) is int:
             if (
                 not is_tensor_array
-                and isinstance(slice_item, int)
                 and x.shape[dim] is not None
                 and x.shape[dim] >= 0
                 and slice_item >= x.shape[dim]
@@ -308,6 +305,13 @@ def parse_index(x, indices):
             step = 1
             end = slice_item + 1 if slice_item != -1 else MAX_INTEGER
             dim += 1
+        elif is_scalar_tensor(slice_item):
+            # not calculate result to reduce call times for slice OP.
+            decrease_axes.append(dim)
+            start = slice_item
+            step = 1
+            end = slice_item + 1
+            dim += 1
         elif isinstance(slice_item, bool):
             # single bool is advanced-indexing
             none_axes.append(dim)

python/paddle/decomposition/decomp.py

@@ -84,7 +84,7 @@ def _prepare_python_api_arguments(op):
     inputs = []
     for x in op.operands():
         input = x.source()
-        if input and input.initialized():
+        if input.initialized():
             prev_op = input.get_defining_op()
             if (
                 isinstance(prev_op, Operation)
@@ -111,7 +111,7 @@ def _check_prim_dynamic(op):
     inputs = []
     for x in op.operands():
         input = x.source()
-        if input and input.initialized():
+        if input.initialized():
             prev_op = input.get_defining_op()
             if (
                 isinstance(prev_op, Operation)

python/paddle/nn/functional/loss.py

@@ -3733,7 +3733,13 @@ def triplet_margin_with_distance_loss(
         swap_dist = distance_function(positive, negative)
         negative_dist = paddle.minimum(negative_dist, swap_dist)
 
-    if not paddle.all(positive_dist > 0) or not paddle.all(negative_dist > 0):
+    if (
+        not isinstance(positive_dist, paddle.pir.Value)
+        and not paddle.all(positive_dist > 0)
+    ) or (
+        not isinstance(negative_dist, paddle.pir.Value)
+        and not paddle.all(negative_dist > 0)
+    ):
         raise ValueError(
             "The positive distance or negative distance should be greater than 0, "
             "The distance functions should be checked."

python/paddle/optimizer/adamw.py

@@ -175,11 +175,11 @@ def __init__(
         assert beta1 is not None
         assert beta2 is not None
         assert epsilon is not None
-        if not 0 <= beta1 < 1:
+        if not isinstance(beta1, Value) and not 0 <= beta1 < 1:
             raise ValueError("Invaild value of beta1, expect beta1 in [0,1).")
-        if not 0 <= beta2 < 1:
+        if not isinstance(beta1, Value) and not 0 <= beta2 < 1:
             raise ValueError("Invaild value of beta2, expect beta2 in [0,1).")
-        if not 0 <= epsilon:
+        if not isinstance(beta1, Value) and not 0 <= epsilon:
             raise ValueError("Invaild value of epsilon, expect epsilon >= 0.")
         if not isinstance(weight_decay, float) and not isinstance(
             weight_decay, (framework.Variable, Value)

python/paddle/pir/math_op_patch.py

@@ -448,6 +448,13 @@ def _float_(self):
             "2. If you want to run it in full graph mode, you need use Value directly, and do not use float(Value)."
         )
 
+    def _bool_(self):
+        raise TypeError(
+            "bool(Value) is not supported in static graph mode. If you are using @to_static, you can try this:\n"
+            "1. If you want to get the value of Value, you can switch to non-fullgraph mode by setting @to_static(full_graph=True).\n"
+            "2. If you want to run it in full graph mode, you need use Value.astype(paddle.bool), and do not use bool(Value)."
+        )
+
     def clone(self):
         """
         Returns a new static Value, which is the clone of the original static
@@ -669,6 +676,7 @@ def value_hash(self):
         ),
         ('__float__', _float_),
         ('__int__', _int_),
+        ('__bool__', _bool_),
     ]
 
     global _already_patch_value

python/paddle/tensor/manipulation.py

@@ -5802,7 +5802,7 @@ def take_along_axis(arr, indices, axis, broadcast=True):
                 )
 
         axis_max_size = arr.shape[axis]
-        if not (indices < axis_max_size).all():
+        if in_dynamic_mode() and not (indices < axis_max_size).all():
             raise RuntimeError(
                 "one of element of indices is out of bounds for dimension {} with size {}".format(
                     axis, axis_max_size
@@ -5958,7 +5958,7 @@ def put_along_axis(
             if elements == 1:  # paddle.pir.Value has no attribute 'size'
                 values = paddle.broadcast_to(values, indices.shape)
         axis_max_size = arr.shape[axis]
-        if not (indices < axis_max_size).all():
+        if in_dynamic_mode() and not (indices < axis_max_size).all():
             raise RuntimeError(
                 "one of element of indices is out of bounds for dimension {} with size {}".format(
                     axis, axis_max_size

test/dygraph_to_static/test_tensor_shape.py

@@ -748,7 +748,7 @@ def _set_test_func(self):
         self.dygraph_func = dyfunc_with_for_1
 
     def _set_expected_op_num(self):
-        self.expected_op_num = 29
+        self.expected_op_num = 27
         self.expected_shape_op_num = 2
         self.expected_slice_op_num = 3
 

test/legacy_test/test_math_op_patch_pir.py

@@ -533,6 +533,8 @@ def test_builtin_type_conversion(self):
                     int(x)
                 with self.assertRaises(TypeError):
                     float(x)
+                with self.assertRaises(TypeError):
+                    bool(x)
 
     def test_math_exists(self):
         with paddle.pir_utils.IrGuard():