diff --git a/paddle/phi/kernels/impl/einsum_impl.h b/paddle/phi/kernels/impl/einsum_impl.h
index 80529c8b669aac9aa04f4883d4465a6fd720bcb1..dafb967ae8ed5c9d8fc227928219fc4c76305f36 100644
--- a/paddle/phi/kernels/impl/einsum_impl.h
+++ b/paddle/phi/kernels/impl/einsum_impl.h
@@ -241,7 +241,7 @@ inline static void InferLabelShape(const std::vector<std::string>& op_labels,
       } else if (labelshape->is_default(c) || (*labelshape)[c] == -1) {
         (*labelshape)[c] = op_dim[dim_ptr];
         dim_ptr++;
-      } else {
+      } else if (op_dim[dim_ptr] != -1) {
         PADDLE_ENFORCE_EQ(
             (*labelshape)[c],
             op_dim[dim_ptr],
diff --git a/python/paddle/fluid/dygraph/base.py b/python/paddle/fluid/dygraph/base.py
index 5101483858d0a21965bc777d8ed95b459e1953a5..82ef806583743e908a2fe5b05d98fe3b9e3ca5ba 100644
--- a/python/paddle/fluid/dygraph/base.py
+++ b/python/paddle/fluid/dygraph/base.py
@@ -27,6 +27,7 @@ from ..data_feeder import convert_dtype
 import warnings
 from ..framework import _get_paddle_place, _in_legacy_dygraph, _in_eager_without_dygraph_check
 import paddle
+import warnings
 
 __all__ = [
     'no_grad', 'no_grad_', 'grad', 'guard', 'enable_dygraph', 'disable_dygraph',
@@ -45,6 +46,20 @@ def in_declarative_mode():
     return _in_declarative_mode_
 
 
+def declarative_unsupport_argument_warning(func_name, input_names, inputs,
+                                           support_values):
+    """
+    Warning if inputs do not elementwisely equals to support_values.
+    It's a utility function for dy2static when dygraph interface have
+    more inputs than static interface such as paddle.grad.
+
+    """
+    for name, inp, sup in zip(input_names, inputs, support_values):
+        if inp != sup:
+            warnings.warn(f"{func_name} has unsupported parameter in jit: " +
+                          f"{name}, jit will discard it")
+
+
 def _switch_to_static_graph_(func):
 
     def __impl__(*args, **kwargs):
@@ -290,6 +305,10 @@ def no_grad(func=None):
         test_layer()
 
     """
+    if in_declarative_mode():
+        warnings.warn(
+            "paddle.no_grad is only supported for inference model, and not supported for training under @to_static."
+        )
     if func is None:
         return _switch_tracer_mode_guard_(is_train=False)
     else:
@@ -428,7 +447,7 @@ def guard(place=None):
                     yield
 
 
-@framework.dygraph_only
+@framework.non_static_only
 def grad(outputs,
          inputs,
          grad_outputs=None,
@@ -563,6 +582,16 @@ def grad(outputs,
             grad_y1 = paddle.to_tensor(3.0)
             print(test_dygraph_grad([grad_y1, grad_value])) # [24.]
 	'''
+    if in_declarative_mode():
+        # In dy2static context, we call static interface `gradients`
+        # to calculate grads.
+        from paddle.static import gradients
+        declarative_unsupport_argument_warning(
+            "paddle.grad",
+            ["retain_graph", "create_grad", "only_inputs", "allow_unused"],
+            [retain_graph, create_graph, only_inputs, allow_unused],
+            [None, False, True, False])
+        return gradients(outputs, inputs, grad_outputs, no_grad_vars)
 
     def check_in_out(in_out_list, name):
         assert in_out_list is not None, "{} should not be None".format(name)
diff --git a/python/paddle/fluid/dygraph/dygraph_to_static/ast_transformer.py b/python/paddle/fluid/dygraph/dygraph_to_static/ast_transformer.py
index 0853b15e6830c4ba5c4d4c14ff5b51d0a14ae696..5ec1dbea504cc6f536d13c34260b1c77e3681366 100644
--- a/python/paddle/fluid/dygraph/dygraph_to_static/ast_transformer.py
+++ b/python/paddle/fluid/dygraph/dygraph_to_static/ast_transformer.py
@@ -102,7 +102,7 @@ class DygraphToStaticAst(BaseTransformer):
             PrintTransformer,  # print statement
             CallTransformer,  # transform call recursively
             CastTransformer,  # type casting statement
-            GradTransformer,  # transform paddle.grad to paddle.gradients
+            #GradTransformer,  # transform paddle.grad to paddle.gradients
             DecoratorTransformer,  # transform decorators to function call
         ]
 
diff --git a/python/paddle/fluid/framework.py b/python/paddle/fluid/framework.py
index 40223d4bcc2a9d8a33a1b7093f28e6f5a081165f..85d525ab5f75cb3d2f2124e72979580cc4a67e8a 100644
--- a/python/paddle/fluid/framework.py
+++ b/python/paddle/fluid/framework.py
@@ -513,6 +513,17 @@ def _dygraph_only_(func):
     return __impl__
 
 
+def _non_static_only_(func):
+
+    def __impl__(*args, **kwargs):
+        from .dygraph.base import in_declarative_mode
+        assert _non_static_mode() or in_declarative_mode(
+        ), "We only support '%s()' in dynamic graph mode, please call 'paddle.disable_static()' to enter dynamic graph mode." % func.__name__
+        return func(*args, **kwargs)
+
+    return __impl__
+
+
 def _static_only_(func):
 
     def __impl__(*args, **kwargs):
@@ -572,6 +583,7 @@ dygraph_not_support = wrap_decorator(_dygraph_not_support_)
 dygraph_only = wrap_decorator(_dygraph_only_)
 static_only = wrap_decorator(_static_only_)
 fake_interface_only = wrap_decorator(_fake_interface_only_)
+non_static_only = wrap_decorator(_non_static_only_)
 
 
 def _dygraph_tracer():