test_imperative_basic.py

# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import unittest

import numpy as np
from test_imperative_base import new_program_scope

import paddle
import paddle.fluid as fluid
import paddle.fluid.dygraph_utils as dygraph_utils
from paddle.fluid import core
from paddle.fluid.dygraph.layer_object_helper import LayerObjectHelper
from paddle.fluid.framework import _in_legacy_dygraph, _test_eager_guard
from paddle.fluid.layer_helper import LayerHelper


class MyLayer(fluid.Layer):
    def __init__(self):
        super().__init__()

    def forward(self, inputs):
        x = fluid.layers.relu(inputs)
        self._x_for_debug = x
        x = paddle.multiply(x, x)
        x = paddle.sum(x)
        return [x]


class MLP(fluid.Layer):
    def __init__(self, input_size):
        super().__init__()
        self._linear1 = paddle.nn.Linear(
            input_size,
            3,
            weight_attr=paddle.ParamAttr(
                initializer=paddle.nn.initializer.Constant(value=0.1)
            ),
            bias_attr=paddle.ParamAttr(
                initializer=paddle.nn.initializer.Constant(value=0.1)
            ),
        )
        self._linear2 = paddle.nn.Linear(
            3,
            4,
            weight_attr=paddle.ParamAttr(
                initializer=paddle.nn.initializer.Constant(value=0.1)
            ),
            bias_attr=paddle.ParamAttr(
                initializer=paddle.nn.initializer.Constant(value=0.1)
            ),
        )

    def forward(self, inputs):
        x = self._linear1(inputs)
        x = self._linear2(x)
        x = paddle.sum(x)
        return x


class SimpleRNNCell(fluid.Layer):
    def __init__(self, step_input_size, hidden_size, output_size, param_attr):
        super().__init__()
        self.step_input_size = step_input_size
        self.hidden_size = hidden_size
        self.output_size = output_size
        self._dtype = core.VarDesc.VarType.FP32
        self.param_attr = param_attr

        i2h_param_shape = [self.step_input_size, self.hidden_size]
        h2h_param_shape = [self.hidden_size, self.hidden_size]
        h2o_param_shape = [self.output_size, self.hidden_size]
        self._i2h_w = None
        self._i2h_w = self.create_parameter(
            attr=self.param_attr,
            shape=i2h_param_shape,
            dtype=self._dtype,
            is_bias=False,
        )
        self._h2h_w = self.create_parameter(
            attr=self.param_attr,
            shape=h2h_param_shape,
            dtype=self._dtype,
            is_bias=False,
        )
        self._h2o_w = self.create_parameter(
            attr=self.param_attr,
            shape=h2o_param_shape,
            dtype=self._dtype,
            is_bias=False,
        )

    def forward(self, input, pre_hidden):
        tmp_i2h = paddle.fluid.layers.nn.mul(input, self._i2h_w)
        tmp_h2h = paddle.fluid.layers.nn.mul(pre_hidden, self._h2h_w)
        hidden = paddle.add(tmp_h2h, tmp_i2h)
        hidden = self._helper.append_activation(hidden, act='tanh')
        out = paddle.fluid.layers.nn.mul(hidden, self._h2o_w)
        softmax_out = paddle.nn.functional.softmax(out)
        reduce_out = paddle.sum(softmax_out)
        return reduce_out, hidden


class SimpleRNN(fluid.Layer):
    def __init__(self):
        super().__init__()
        self.seq_len = 4
        self._cell = SimpleRNNCell(
            3,
            3,
            3,
            fluid.ParamAttr(initializer=fluid.initializer.Constant(value=0.1)),
        )

    def forward(self, inputs):
        outs = list()
        pre_hiddens = list()

        init_hidden = self.create_parameter(
            attr=fluid.ParamAttr(
                initializer=fluid.initializer.Constant(value=0.1)
            ),
            shape=[1, 3],
            dtype='float32',
            is_bias=False,
        )
        pre_hidden = init_hidden
        for i in range(self.seq_len):
            input = paddle.slice(inputs, axes=[1], starts=[i], ends=[i + 1])
            input = paddle.reshape(input, shape=[1, 3])
            out_softmax, pre_hidden = self._cell(input, pre_hidden)
            outs.append(out_softmax)

        return outs, pre_hiddens


class TestImperative(unittest.TestCase):
    def functional_dygraph_context(self):
        self.assertFalse(fluid.dygraph.enabled())
        fluid.enable_dygraph()
        self.assertTrue(fluid.dygraph.enabled())
        np_inp = np.array([[1.0, 2.0], [3.0, 4.0]], dtype=np.float32)
        var_inp = paddle.to_tensor(np_inp)
        mlp = MLP(input_size=2)
        out = mlp(var_inp)
        dy_out1 = out.numpy()
        out.backward()
        dy_grad1 = mlp._linear1.weight.gradient()
        fluid.disable_dygraph()
        self.assertFalse(fluid.dygraph.enabled())
        with fluid.dygraph.guard():
            self.assertTrue(fluid.dygraph.enabled())
            var_inp = paddle.to_tensor(np_inp)
            mlp = MLP(input_size=2)
            out = mlp(var_inp)
            dy_out2 = out.numpy()
            out.backward()
            dy_grad2 = mlp._linear1.weight.gradient()
        self.assertFalse(fluid.dygraph.enabled())
        np.testing.assert_array_equal(dy_out1, dy_out2)
        np.testing.assert_array_equal(dy_grad1, dy_grad2)

    def test_functional_dygraph_context(self):
        with _test_eager_guard():
            self.functional_dygraph_context()
        self.functional_dygraph_context()

    def functional_paddle_imperative_dygraph_context(self):
        self.assertFalse(paddle.in_dynamic_mode())
        paddle.disable_static()
        self.assertTrue(paddle.in_dynamic_mode())
        np_inp = np.array([[1.0, 2.0], [3.0, 4.0]], dtype=np.float32)
        var_inp = paddle.to_tensor(np_inp)
        mlp = MLP(input_size=2)
        out = mlp(var_inp)
        dy_out1 = out.numpy()
        out.backward()
        dy_grad1 = mlp._linear1.weight.gradient()
        paddle.enable_static()
        self.assertFalse(paddle.in_dynamic_mode())
        paddle.disable_static()
        self.assertTrue(paddle.in_dynamic_mode())
        var_inp = paddle.to_tensor(np_inp)
        mlp = MLP(input_size=2)
        out = mlp(var_inp)
        dy_out2 = out.numpy()
        out.backward()
        dy_grad2 = mlp._linear1.weight.gradient()
        paddle.enable_static()
        self.assertFalse(paddle.in_dynamic_mode())
        np.testing.assert_array_equal(dy_out1, dy_out2)
        np.testing.assert_array_equal(dy_grad1, dy_grad2)

    def test_functional_paddle_imperative_dygraph_context(self):
        with _test_eager_guard():
            self.functional_paddle_imperative_dygraph_context()
        self.functional_paddle_imperative_dygraph_context()

    def func_isinstance(self):
        var = fluid.layers.data(shape=[1], name='x', dtype='float32')
        self.assertTrue(isinstance(var, fluid.Variable))
        with fluid.dygraph.guard():
            if not _in_legacy_dygraph():
                var_base = paddle.to_tensor(np.array([3, 4, 5]))
                self.assertTrue(isinstance(var_base, core.eager.Tensor))
            else:
                var_base = paddle.to_tensor(np.array([3, 4, 5]))
                self.assertTrue(isinstance(var_base, core.VarBase))
                self.assertTrue(isinstance(var_base, fluid.Variable))

    def test_isinstance(self):
        with _test_eager_guard():
            self.func_isinstance()
        self.func_isinstance()

    def func_create_varbase(self):
        x = np.ones([2, 2], np.float32)
        y = np.zeros([3, 3], np.float32)
        t = fluid.Tensor()
        t.set(x, fluid.CPUPlace())
        if not _in_legacy_dygraph():
            egr_tmp = fluid.core.eager.Tensor(
                value=x, place=fluid.core.CPUPlace()
            )
            egr_tmp2 = fluid.core.eager.Tensor(y, fluid.core.CPUPlace())
            egr_tmp3 = paddle.to_tensor(x)
            egr_tmp4 = fluid.core.eager.Tensor(y)
            egr_tmp5 = fluid.core.eager.Tensor(value=x)
            egr_tmp6 = fluid.core.eager.Tensor(t)

            np.testing.assert_array_equal(x, egr_tmp.numpy())
            np.testing.assert_array_equal(y, egr_tmp2.numpy())
            np.testing.assert_array_equal(x, egr_tmp3.numpy())
            np.testing.assert_array_equal(y, egr_tmp4.numpy())
            np.testing.assert_array_equal(x, egr_tmp5.numpy())
            np.testing.assert_array_equal(x, egr_tmp6.numpy())
        else:
            tmp = fluid.core.VarBase(value=x, place=fluid.core.CPUPlace())
            tmp2 = fluid.core.VarBase(y, fluid.core.CPUPlace())
            tmp3 = paddle.to_tensor(x)
            tmp4 = fluid.core.VarBase(y)
            tmp5 = fluid.core.VarBase(value=x)
            tmp6 = fluid.core.VarBase(t)

            np.testing.assert_array_equal(x, tmp.numpy())
            np.testing.assert_array_equal(y, tmp2.numpy())
            np.testing.assert_array_equal(x, tmp3.numpy())
            np.testing.assert_array_equal(y, tmp4.numpy())
            np.testing.assert_array_equal(x, tmp5.numpy())
            np.testing.assert_array_equal(x, tmp6.numpy())

    def test_create_varbase(self):
        with fluid.dygraph.guard():
            with _test_eager_guard():
                self.func_create_varbase()
            self.func_create_varbase()

    def test_no_grad_guard(self):
        data = np.array([[2, 3], [4, 5]]).astype('float32')
        with fluid.dygraph.guard():
            l0 = paddle.nn.Linear(2, 2)
            self.assertIsNone(l0.weight._grad_ivar())
            l1 = paddle.nn.Linear(2, 2)
            with fluid.dygraph.no_grad():
                self.assertTrue(l1.weight.stop_gradient is False)
                tmp = l1.weight * 2
                self.assertTrue(tmp.stop_gradient)
            x = paddle.to_tensor(data)
            y = paddle.add(l0(x), tmp)
            o = l1(y)
            o.backward()

            self.assertIsNone(tmp._grad_ivar())
            self.assertIsNotNone(l0.weight._grad_ivar())

    def test_paddle_imperative_no_grad_guard(self):
        data = np.array([[2, 3], [4, 5]]).astype('float32')
        with fluid.dygraph.guard():
            l0 = paddle.nn.Linear(2, 2)
            self.assertIsNone(l0.weight._grad_ivar())
            l1 = paddle.nn.Linear(2, 2)
            with paddle.no_grad():
                self.assertTrue(l1.weight.stop_gradient is False)
                tmp = l1.weight * 2
                self.assertTrue(tmp.stop_gradient)
            x = paddle.to_tensor(data)
            y = paddle.add(l0(x), tmp)
            o = l1(y)
            o.backward()

            self.assertIsNone(tmp._grad_ivar())
            self.assertIsNotNone(l0.weight._grad_ivar())

    def test_paddle_imperative_set_grad_enabled(self):
        data = np.array([[2, 3], [4, 5]]).astype('float32')
        with fluid.dygraph.guard():
            l0 = paddle.nn.Linear(2, 2)
            self.assertIsNone(l0.weight._grad_ivar())
            l1 = paddle.nn.Linear(2, 2)
            with paddle.set_grad_enabled(False):
                self.assertTrue(l1.weight.stop_gradient is False)
                tmp = l1.weight * 2
                with paddle.set_grad_enabled(True):
                    tmp2 = l1.weight * 2
                self.assertTrue(tmp.stop_gradient)
                self.assertTrue(tmp2.stop_gradient is False)
            x = paddle.to_tensor(data)
            y = paddle.add(l0(x), tmp2)
            o = l1(y)
            o.backward()

            self.assertIsNone(tmp._grad_ivar())
            self.assertIsNotNone(tmp2._grad_ivar())
            self.assertIsNotNone(l0.weight._grad_ivar())

    def test_paddle_imperative_is_grad_enabled(self):
        with fluid.dygraph.guard():
            with paddle.set_grad_enabled(False):
                self.assertTrue(paddle.is_grad_enabled() is False)
                with paddle.set_grad_enabled(True):
                    self.assertTrue(paddle.is_grad_enabled())

    def func_sum_op(self):
        x = np.ones([2, 2], np.float32)
        with fluid.dygraph.guard():
            inputs = []
            for _ in range(10):
                tmp = paddle.to_tensor(x)
                tmp.stop_gradient = False
                inputs.append(tmp)
            ret = paddle.add_n(inputs)
            loss = paddle.sum(ret)
            loss.backward()
        with fluid.dygraph.guard():
            inputs2 = []
            for _ in range(10):
                tmp = paddle.to_tensor(x)
                tmp.stop_gradient = False
                inputs2.append(tmp)
            ret2 = paddle.add_n(inputs2)
            loss2 = paddle.sum(ret2)
            fluid.set_flags({'FLAGS_sort_sum_gradient': True})
            loss2.backward()

            np.testing.assert_allclose(ret.numpy(), x * 10, rtol=1e-05)
            np.testing.assert_allclose(inputs[0].gradient(), x, rtol=1e-05)
            np.testing.assert_allclose(ret2.numpy(), x * 10, rtol=1e-05)
            a = inputs2[0].gradient()
            np.testing.assert_allclose(inputs2[0].gradient(), x, rtol=1e-05)

    def test_sum_op(self):
        with _test_eager_guard():
            self.func_sum_op()
        self.func_sum_op()

    def func_empty_var(self):
        with fluid.dygraph.guard():
            cur_program = fluid.Program()
            cur_block = cur_program.current_block()
            # Normally, we don't allow tensor with -1 shape being created in dygraph mode, this test is not good.
            if _in_legacy_dygraph():
                new_variable = cur_block.create_var(
                    name="X", shape=[-1, 23, 48], dtype='float32'
                )
            else:
                new_variable = cur_block.create_var(
                    name="X", shape=[1, 23, 48], dtype='float32'
                )
            try:
                new_variable.numpy()
            except Exception as e:
                assert type(e) == ValueError

            try:
                new_variable.backward()
            except Exception as e:
                assert type(e) == core.EnforceNotMet
            try:
                new_variable.clear_gradient()
            except Exception as e:
                assert type(e) == core.EnforceNotMet

    def test_empty_var(self):
        with _test_eager_guard():
            self.func_empty_var()
        self.func_empty_var()

    def func_empty_grad(self):
        with fluid.dygraph.guard():
            x = np.ones([2, 2], np.float32)
            new_var = paddle.to_tensor(x)
            self.assertIsNone(new_var.gradient())
            try:
                new_var.clear_gradient()
            except Exception as e:
                assert type(e) == core.EnforceNotMet

        with fluid.dygraph.guard():
            cur_program = fluid.Program()
            cur_block = cur_program.current_block()
            # Normally, we don't allow tensor with -1 shape being created in dygraph mode, this test is not good.
            if _in_legacy_dygraph():
                new_variable = cur_block.create_var(
                    name="X", shape=[-1, 23, 48], dtype='float32'
                )
            else:
                new_variable = cur_block.create_var(
                    name="X", shape=[1, 23, 48], dtype='float32'
                )
            try:
                new_variable.gradient()
            except Exception as e:
                assert type(e) == ValueError

    def test_empty_grad(self):
        with _test_eager_guard():
            self.func_empty_grad()
        self.func_empty_grad()

    def func_set_persistable(self):
        with fluid.dygraph.guard():
            x = np.ones([2, 2], np.float32)
            new_var = paddle.to_tensor(x)
            self.assertFalse(new_var.persistable)
            new_var.persistable = True
            self.assertTrue(new_var.persistable)

    def test_set_persistable(self):
        with _test_eager_guard():
            self.func_set_persistable()
        self.func_set_persistable()

    def func_layer(self):
        with fluid.dygraph.guard():
            l = fluid.Layer("l")
            self.assertRaises(NotImplementedError, l.forward, [])

    def test_layer(self):
        with _test_eager_guard():
            self.func_layer()
        self.func_layer()

    def func_layer_in_out(self):
        np_inp = np.array([1.0, 2.0, -1.0], dtype=np.float32)
        with fluid.dygraph.guard():
            var_inp = paddle.to_tensor(np_inp)
            var_inp.stop_gradient = False
            l = MyLayer()
            x = l(var_inp)[0]
            self.assertIsNotNone(x)
            dy_out = x.numpy()
            x.backward()
            dy_grad = l._x_for_debug.gradient()

        with fluid.dygraph.guard():
            var_inp2 = paddle.to_tensor(np_inp)
            var_inp2.stop_gradient = False
            l2 = MyLayer()
            x2 = l2(var_inp2)[0]
            self.assertIsNotNone(x2)
            dy_out2 = x2.numpy()
            fluid.set_flags({'FLAGS_sort_sum_gradient': True})
            x2.backward()
            dy_grad2 = l2._x_for_debug.gradient()

        with new_program_scope():
            inp = fluid.layers.data(
                name="inp", shape=[3], append_batch_size=False
            )
            l = MyLayer()
            x = l(inp)[0]
            param_grads = fluid.backward.append_backward(
                x, parameter_list=[l._x_for_debug.name]
            )[0]
            exe = fluid.Executor(
                fluid.CPUPlace()
                if not core.is_compiled_with_cuda()
                else fluid.CUDAPlace(0)
            )

            static_out, static_grad = exe.run(
                feed={inp.name: np_inp},
                fetch_list=[x.name, param_grads[1].name],
            )

        np.testing.assert_array_equal(dy_out, static_out)
        np.testing.assert_array_equal(dy_grad, static_grad)
        np.testing.assert_array_equal(dy_out2, static_out)
        np.testing.assert_array_equal(dy_grad2, static_grad)

    def test_layer_in_out(self):
        fluid.set_flags({"FLAGS_retain_grad_for_all_tensor": True})
        with _test_eager_guard():
            self.func_layer_in_out()
        self.func_layer_in_out()
        fluid.set_flags({"FLAGS_retain_grad_for_all_tensor": False})

    def func_mlp(self):
        np_inp = np.array([[1.0, 2.0], [3.0, 4.0]], dtype=np.float32)
        with fluid.dygraph.guard():
            var_inp = paddle.to_tensor(np_inp)
            mlp = MLP(input_size=2)
            out = mlp(var_inp)
            dy_out = out.numpy()
            out.backward()
            dy_grad = mlp._linear1.weight.gradient()

        with fluid.dygraph.guard():
            var_inp2 = paddle.to_tensor(np_inp)
            mlp2 = MLP(input_size=2)
            out2 = mlp2(var_inp2)
            dy_out2 = out2.numpy()
            fluid.set_flags({'FLAGS_sort_sum_gradient': True})
            out2.backward()
            dy_grad2 = mlp2._linear1.weight.gradient()

        with new_program_scope():
            inp = fluid.layers.data(
                name="inp", shape=[2, 2], append_batch_size=False
            )
            mlp = MLP(input_size=2)
            out = mlp(inp)
            param_grads = fluid.backward.append_backward(
                out, parameter_list=[mlp._linear1.weight.name]
            )[0]
            exe = fluid.Executor(
                fluid.CPUPlace()
                if not core.is_compiled_with_cuda()
                else fluid.CUDAPlace(0)
            )
            exe.run(fluid.default_startup_program())

            static_out, static_grad = exe.run(
                feed={inp.name: np_inp},
                fetch_list=[out.name, param_grads[1].name],
            )

        np.testing.assert_allclose(dy_out, static_out, rtol=1e-05)
        np.testing.assert_allclose(dy_grad, static_grad, rtol=1e-05)
        np.testing.assert_allclose(dy_out2, static_out, rtol=1e-05)
        np.testing.assert_allclose(dy_grad2, static_grad, rtol=1e-05)

        params = mlp.parameters(True)
        self.assertEqual("linear_0.w_0", params[0].name)
        self.assertEqual("linear_0.b_0", params[1].name)
        self.assertEqual("linear_1.w_0", params[2].name)
        self.assertEqual("linear_1.b_0", params[3].name)
        self.assertEqual(len(params), 4)

        sublayers = mlp.sublayers()
        self.assertEqual(mlp._linear1, sublayers[0])
        self.assertEqual(mlp._linear2, sublayers[1])
        self.assertEqual(len(sublayers), 2)

    def test_mlp(self):
        with _test_eager_guard():
            self.func_mlp()
        self.func_mlp()

    def test_gradient_accumulation(self):
        def test_single_api(sort_sum_gradient):
            fluid.set_flags({'FLAGS_sort_sum_gradient': sort_sum_gradient})
            x = paddle.to_tensor(5.0, stop_gradient=False)
            for i in range(10):
                y = paddle.pow(x, 4.0)
                y.backward()
                self.assertEqual(x.grad.numpy(), (i + 1) * 500)
            x.clear_gradient()
            self.assertEqual(x.grad.numpy(), 0.0)
            for i in range(10):
                y = paddle.pow(x, 4.0)
                y.backward()
                self.assertEqual(x.grad.numpy(), (i + 1) * 500)
            x.clear_grad()
            self.assertEqual(x.grad.numpy(), 0.0)

        def test_simple_net(sort_sum_gradient):
            fluid.set_flags({'FLAGS_sort_sum_gradient': sort_sum_gradient})
            x = paddle.to_tensor(5.0, stop_gradient=False)
            y = paddle.to_tensor(2.0, stop_gradient=False)
            z = paddle.to_tensor(3.0, stop_gradient=False)

            def fun(x, y, z):
                loss1 = x * x * y
                loss2 = x * z
                loss1.backward(retain_graph=True)
                loss2.backward(retain_graph=True)
                np.testing.assert_array_equal(x.grad.numpy(), [23.0])
                np.testing.assert_array_equal(y.grad.numpy(), [25.0])
                np.testing.assert_array_equal(z.grad.numpy(), [5.0])
                x.clear_grad()
                y.clear_grad()
                z.clear_grad()

                dx = paddle.grad([loss1], x, create_graph=True)[0]
                loss = loss1 + loss2 + dx
                # loss = x*x*y + x*z + 2*x*y
                return loss

            loss = fun(x, y, z)
            loss.backward(retain_graph=True)
            # x.grad = 2*x*y + z + 2*y = 27
            np.testing.assert_array_equal(x.grad.numpy(), [27])

            loss.backward(retain_graph=True)
            np.testing.assert_array_equal(x.grad.numpy(), [54])

            loss.backward()
            np.testing.assert_array_equal(x.grad.numpy(), [81])

            with self.assertRaises(RuntimeError):
                loss.backward()

            loss1 = x * x * y
            loss2 = x * z
            dx = paddle.grad([loss1], x, create_graph=True)[0]
            loss = loss1 + loss2 + dx
            loss.backward()
            np.testing.assert_array_equal(dx.grad.numpy(), [1])
            np.testing.assert_array_equal(x.grad.numpy(), [108])

        def test_mlp(sort_sum_gradient):
            fluid.set_flags({'FLAGS_sort_sum_gradient': sort_sum_gradient})
            input_size = 5
            paddle.seed(1)
            mlp1 = MLP(input_size=input_size)
            # generate the gradient of each step
            mlp2 = MLP(input_size=input_size)

            expected_weight1_grad = 0.0
            expected_bias1_grad = 0.0
            expected_weight2_grad = 0.0
            expected_bias2_grad = 0.0

            for batch_id in range(100):
                x = paddle.uniform([10, input_size])
                detach_x = x.detach()
                clear_loss = mlp2(detach_x)
                clear_loss.backward()
                expected_weight1_grad = (
                    expected_weight1_grad + mlp2._linear1.weight.grad.numpy()
                )
                expected_bias1_grad = (
                    expected_bias1_grad + mlp2._linear1.bias.grad.numpy()
                )
                expected_weight2_grad = (
                    expected_weight2_grad + mlp2._linear2.weight.grad.numpy()
                )
                expected_bias2_grad = (
                    expected_bias2_grad + mlp2._linear2.bias.grad.numpy()
                )

                loss = mlp1(x)
                loss.backward()

                np.testing.assert_array_equal(loss.grad.numpy(), [1])
                np.testing.assert_allclose(
                    mlp1._linear1.weight.grad.numpy(),
                    expected_weight1_grad,
                    rtol=1e-05,
                )
                np.testing.assert_allclose(
                    mlp1._linear1.bias.grad.numpy(),
                    expected_bias1_grad,
                    rtol=1e-05,
                )
                np.testing.assert_allclose(
                    mlp1._linear2.weight.grad.numpy(),
                    expected_weight2_grad,
                    rtol=1e-05,
                )
                np.testing.assert_allclose(
                    mlp1._linear2.bias.grad.numpy(),
                    expected_bias2_grad,
                    rtol=1e-05,
                )

                mlp2.clear_gradients()
                np.testing.assert_array_equal(clear_loss.grad.numpy(), [1])
                if ((batch_id + 1) % 10) % 2 == 0:
                    mlp1.clear_gradients()
                    expected_weight1_grad = 0.0
                    expected_bias1_grad = 0.0
                    expected_weight2_grad = 0.0
                    expected_bias2_grad = 0.0
                elif ((batch_id + 1) % 10) % 2 == 1:
                    mlp1.clear_gradients()
                    mlp1._linear1.weight._set_grad_ivar(
                        paddle.ones([input_size, 3])
                    )
                    mlp1._linear2.weight._set_grad_ivar(paddle.ones([3, 4]))
                    expected_weight1_grad = 1.0
                    expected_bias1_grad = 0.0
                    expected_weight2_grad = 1.0
                    expected_bias2_grad = 0.0

        with fluid.dygraph.guard():
            test_single_api(False)
            test_single_api(True)
            test_simple_net(False)
            test_simple_net(True)
            test_mlp(False)
            test_mlp(True)

    def func_dygraph_vs_static(self):
        np_inp1 = np.random.rand(4, 3, 3)
        np_inp2 = np.random.rand(4, 3, 3)

        # dynamic graph
        with fluid.dygraph.guard():
            inp1 = paddle.to_tensor(np_inp1)
            inp2 = paddle.to_tensor(np_inp2)
            if np.sum(np_inp1) < np.sum(np_inp2):
                x = paddle.add(inp1, inp2)
            else:
                x = paddle.subtract(inp1, inp2)
            dygraph_result = x.numpy()

        # static graph
        with new_program_scope():
            inp_data1 = fluid.layers.data(
                name='inp1', shape=[3, 3], dtype=np.float32
            )
            inp_data2 = fluid.layers.data(
                name='inp2', shape=[3, 3], dtype=np.float32
            )

            a = paddle.expand(
                paddle.reshape(paddle.sum(inp_data1), [1, 1]),
                [4, -1],
            )
            b = paddle.expand(
                paddle.reshape(paddle.sum(inp_data2), [1, 1]),
                [4, -1],
            )
            cond = paddle.less_than(x=a, y=b)

            ie = fluid.layers.IfElse(cond)
            with ie.true_block():
                d1 = ie.input(inp_data1)
                d2 = ie.input(inp_data2)
                d3 = paddle.add(d1, d2)
                ie.output(d3)

            with ie.false_block():
                d1 = ie.input(inp_data1)
                d2 = ie.input(inp_data2)
                d3 = paddle.subtract(d1, d2)
                ie.output(d3)
            out = ie()

            exe = fluid.Executor(
                fluid.CPUPlace()
                if not core.is_compiled_with_cuda()
                else fluid.CUDAPlace(0)
            )
            static_result = exe.run(
                fluid.default_main_program(),
                feed={'inp1': np_inp1, 'inp2': np_inp2},
                fetch_list=out,
            )[0]
        np.testing.assert_allclose(dygraph_result, static_result, rtol=1e-05)

    def test_dygraph_vs_static(self):
        with _test_eager_guard():
            self.func_dygraph_vs_static()
        self.func_dygraph_vs_static()

    def func_rnn(self):
        np_inp = np.array(
            [
                [1.0, 2.0, 3.0],
                [4.0, 5.0, 6.0],
                [7.0, 8.0, 9.0],
                [10.0, 11.0, 12.0],
            ]
        )
        np_inp = np_inp.reshape((1, 4, 3))
        np_inp = np_inp.astype(np.float32)
        with fluid.dygraph.guard():
            var_inp = paddle.to_tensor(np_inp)
            var_inp = paddle.reshape(var_inp, shape=[1, 4, 3])
            simple_rnn = SimpleRNN()
            outs, pre_hiddens = simple_rnn.forward(var_inp)
            dy_out = outs[3].numpy()
            outs[3].backward()
            dy_grad_h2o = simple_rnn._cell._h2o_w.gradient()
            dy_grad_h2h = simple_rnn._cell._h2h_w.gradient()
            dy_grad_i2h = simple_rnn._cell._i2h_w.gradient()

        with fluid.dygraph.guard():
            var_inp2 = paddle.to_tensor(np_inp)
            var_inp2 = paddle.reshape(var_inp2, shape=[1, 4, 3])
            simple_rnn2 = SimpleRNN()
            outs2, pre_hiddens2 = simple_rnn2.forward(var_inp2)
            dy_out2 = outs2[3].numpy()
            fluid.set_flags({'FLAGS_sort_sum_gradient': True})
            outs2[3].backward()
            dy_grad_h2o2 = simple_rnn2._cell._h2o_w.gradient()
            dy_grad_h2h2 = simple_rnn2._cell._h2h_w.gradient()
            dy_grad_i2h2 = simple_rnn2._cell._i2h_w.gradient()

        with new_program_scope():
            inp = fluid.layers.data(
                name="inp", shape=[1, 4, 3], append_batch_size=False
            )
            simple_rnn = SimpleRNN()
            outs, pre_hiddens = simple_rnn(inp)
            param_grads = fluid.backward.append_backward(outs[3])
            exe = fluid.Executor(fluid.CPUPlace())
            exe.run(fluid.default_startup_program())
            (
                static_out,
                static_grad_h2o,
                static_grad_h2h,
                static_grad_i2h,
            ) = exe.run(
                feed={inp.name: np_inp},
                fetch_list=[
                    outs[3].name,
                    param_grads[0][1].name,
                    param_grads[1][1].name,
                    param_grads[2][1].name,
                ],
            )

        np.testing.assert_array_equal(dy_out, static_out)
        np.testing.assert_array_equal(dy_grad_h2o, static_grad_h2o)
        np.testing.assert_array_equal(dy_grad_h2h, static_grad_h2h)
        np.testing.assert_array_equal(dy_grad_i2h, static_grad_i2h)
        np.testing.assert_array_equal(dy_out2, static_out)
        np.testing.assert_array_equal(dy_grad_h2o2, static_grad_h2o)
        np.testing.assert_array_equal(dy_grad_h2h2, static_grad_h2h)
        np.testing.assert_array_equal(dy_grad_i2h2, static_grad_i2h)

    def test_rnn(self):
        with _test_eager_guard():
            self.func_rnn()
        self.func_rnn()

    def func_layer_attrs(self):
        layer = fluid.dygraph.Layer("test")
        layer.test_attr = 1
        self.assertFalse(hasattr(layer, "whatever"))
        self.assertTrue(hasattr(layer, "test_attr"))
        self.assertEqual(layer.test_attr, 1)

        my_layer = MyLayer()
        my_layer.w1 = my_layer.create_parameter([3, 3])
        my_layer.add_parameter('w2', None)
        self.assertEqual(len(my_layer.parameters()), 1)
        self.assertRaises(TypeError, my_layer.__setattr__, 'w1', 'str')
        my_layer.w1 = None
        self.assertEqual(len(my_layer.parameters()), 0)
        my_layer.l1 = paddle.nn.Linear(3, 3)
        self.assertEqual(len(my_layer.sublayers()), 1)
        self.assertRaises(TypeError, my_layer.__setattr__, 'l1', 'str')
        my_layer.l1 = None
        self.assertEqual(len(my_layer.sublayers()), 0)

    def test_layer_attrs(self):
        with _test_eager_guard():
            self.func_layer_attrs()
        self.func_layer_attrs()


class TestDygraphUtils(unittest.TestCase):
    def func_append_activation_in_dygraph_exception(self):
        with new_program_scope():
            np_inp = np.random.random(size=(10, 20, 30)).astype(np.float32)
            a = fluid.layers.data("a", [10, 20])
            func = dygraph_utils._append_activation_in_dygraph
            self.assertRaises(AssertionError, func, a, act="sigmoid")

    def test_append_activation_in_dygraph_exception(self):
        with _test_eager_guard():
            self.func_append_activation_in_dygraph_exception()
        self.func_append_activation_in_dygraph_exception()

    def func_append_activation_in_dygraph1(self):
        a_np = np.random.random(size=(10, 20, 30)).astype(np.float32)
        func = dygraph_utils._append_activation_in_dygraph
        with fluid.dygraph.guard():
            a = paddle.to_tensor(a_np)
            res1 = func(a, act="hard_sigmoid")
            res2 = paddle.nn.functional.hardsigmoid(a, slope=0.2)
            np.testing.assert_array_equal(res1.numpy(), res2.numpy())

    def test_append_activation_in_dygraph1(self):
        with _test_eager_guard():
            self.func_append_activation_in_dygraph1()
        self.func_append_activation_in_dygraph1()

    def func_append_activation_in_dygraph2(self):
        a_np = np.random.random(size=(10, 20, 30)).astype(np.float32)
        func = dygraph_utils._append_activation_in_dygraph
        with fluid.dygraph.guard():
            a = paddle.to_tensor(a_np)
            res1 = func(a, act="sigmoid", use_mkldnn=True, use_cudnn=True)
            res2 = paddle.nn.functional.sigmoid(a)
            np.testing.assert_allclose(res1.numpy(), res2.numpy(), rtol=1e-05)

    def test_append_activation_in_dygraph2(self):
        with _test_eager_guard():
            self.func_append_activation_in_dygraph2()
        self.func_append_activation_in_dygraph2()

    def func_append_activation_in_dygraph3(self):
        a_np = np.random.random(size=(10, 20, 30)).astype(np.float32)
        helper = LayerObjectHelper(fluid.unique_name.generate("test"))
        func = helper.append_activation
        with fluid.dygraph.guard():
            a = paddle.to_tensor(a_np)
            res1 = func(a, act="sigmoid", use_cudnn=True)
            res2 = paddle.nn.functional.sigmoid(a)
            np.testing.assert_array_equal(res1.numpy(), res2.numpy())

    def test_append_activation_in_dygraph3(self):
        with _test_eager_guard():
            self.func_append_activation_in_dygraph3()
        self.func_append_activation_in_dygraph3()

    def func_append_activation_in_dygraph_use_mkldnn(self):
        a_np = np.random.uniform(-2, 2, (10, 20, 30)).astype(np.float32)
        helper = LayerHelper(
            fluid.unique_name.generate("test"), act="relu", use_mkldnn=True
        )
        func = helper.append_activation
        with fluid.dygraph.guard():
            a = paddle.to_tensor(a_np)
            res1 = func(a)
            res2 = fluid.layers.relu(a)
            np.testing.assert_array_equal(res1.numpy(), res2.numpy())

    def test_append_activation_in_dygraph_use_mkldnn(self):
        with _test_eager_guard():
            self.func_append_activation_in_dygraph_use_mkldnn()
        self.func_append_activation_in_dygraph_use_mkldnn()

    def func_append_activation_in_dygraph_global_use_mkldnn(self):
        a_np = np.random.uniform(-2, 2, (10, 20, 30)).astype(np.float32)
        helper = LayerHelper(fluid.unique_name.generate("test"), act="relu")
        func = helper.append_activation
        with fluid.dygraph.guard(fluid.core.CPUPlace()):
            a = paddle.to_tensor(a_np)
            fluid.set_flags({'FLAGS_use_mkldnn': True})
            try:
                res1 = func(a)
            finally:
                fluid.set_flags({'FLAGS_use_mkldnn': False})
            res2 = fluid.layers.relu(a)
        np.testing.assert_array_equal(res1.numpy(), res2.numpy())

    def test_append_activation_in_dygraph_global_use_mkldnn(self):
        with _test_eager_guard():
            self.func_append_activation_in_dygraph_global_use_mkldnn()
        self.func_append_activation_in_dygraph_global_use_mkldnn()

    def func_append_bias_in_dygraph_exception(self):
        with new_program_scope():
            np_inp = np.random.random(size=(10, 20, 30)).astype(np.float32)
            a = fluid.layers.data("a", [10, 20])
            func = dygraph_utils._append_bias_in_dygraph
            self.assertRaises(AssertionError, func, a)

    def test_append_bias_in_dygraph_exception(self):
        with _test_eager_guard():
            self.func_append_bias_in_dygraph_exception()
        self.func_append_bias_in_dygraph_exception()

    def func_append_bias_in_dygraph(self):
        a_np = np.random.random(size=(10, 20, 30)).astype(np.float32)
        func = dygraph_utils._append_bias_in_dygraph
        with fluid.dygraph.guard():
            a = paddle.to_tensor(a_np)
            res1 = func(a, bias=a)
            res2 = paddle.add(a, a)
            np.testing.assert_array_equal(res1.numpy(), res2.numpy())

    def test_append_bias_in_dygraph(self):
        with _test_eager_guard():
            self.func_append_bias_in_dygraph()
        self.func_append_bias_in_dygraph()


class TestDygraphGuardWithError(unittest.TestCase):
    def func_without_guard(self):
        with fluid.dygraph.guard():
            x = paddle.to_tensor(np.zeros([10, 10]))
        with self.assertRaisesRegexp(
            TypeError, "Please use `with fluid.dygraph.guard()"
        ):
            y = fluid.layers.matmul(x, x)

    def test_without_guard(self):
        with _test_eager_guard():
            self.func_without_guard()
        self.func_without_guard()


class TestMetaclass(unittest.TestCase):
    def func_metaclass(self):
        self.assertEqual(type(MyLayer).__name__, 'type')
        self.assertNotEqual(type(MyLayer).__name__, 'pybind11_type')
        if not _in_legacy_dygraph():
            self.assertEqual(
                type(paddle.fluid.core.eager.Tensor).__name__, 'type'
            )
        else:
            self.assertEqual(
                type(paddle.fluid.core.VarBase).__name__, 'pybind11_type'
            )

    def test_metaclass(self):
        with _test_eager_guard():
            self.func_metaclass()
        self.func_metaclass()


if __name__ == '__main__':
    paddle.enable_static()
    unittest.main()