1 /**
2 * Copyright 2020 Huawei Technologies Co., Ltd
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16 #include <vector>
17 #include <memory>
18 #include "common/common_test.h"
19 #include "ops/grad/pooling_grad.h"
20 #include "ir/dtype/type.h"
21 #include "ir/value.h"
22 #include "abstract/dshape.h"
23 #include "utils/tensor_construct_utils.h"
24
25 namespace mindspore {
26 namespace ops {
27 class TestPoolingGrad : public UT::Common {
28 public:
TestPoolingGrad()29 TestPoolingGrad() {}
SetUp()30 void SetUp() {}
TearDown()31 void TearDown() {}
32 };
33
TEST_F(TestPoolingGrad,test_ops_pooling_grad1)34 TEST_F(TestPoolingGrad, test_ops_pooling_grad1) {
35 auto pooling_grad = std::make_shared<PoolingGrad>();
36 pooling_grad->Init(MAX_POOLING, std::vector<int64_t>{1, 1}, std::vector<int64_t>{1, 1}, VALID,
37 std::vector<int64_t>{1, 1, 1, 1}, FLOOR, NCHW, false);
38 EXPECT_EQ(pooling_grad->get_pool_mode(), MAX_POOLING);
39 // EXPECT_EQ(pooling_grad->get_window(), std::vector<int64_t>{1, 1});
40 EXPECT_EQ(pooling_grad->get_pad_mode(), VALID);
41 // EXPECT_EQ(pooling_grad->get_stride(), std::vector<int64_t>{1, 1});
42 // EXPECT_EQ(pooling_grad->get_pad_list(), std::vector<int64_t>{1, 1, 1, 1});
43 EXPECT_EQ(pooling_grad->get_round_mode(), FLOOR);
44 EXPECT_EQ(pooling_grad->get_format(), NCHW);
45 EXPECT_EQ(pooling_grad->get_global(), false);
46 auto input0 = TensorConstructUtils::CreateOnesTensor(kNumberTypeFloat32, std::vector<int64_t>{1});
47 auto input1 = TensorConstructUtils::CreateOnesTensor(kNumberTypeFloat32, std::vector<int64_t>{1});
48 auto input2 = TensorConstructUtils::CreateOnesTensor(kNumberTypeFloat32, std::vector<int64_t>{3, 3});
49 MS_EXCEPTION_IF_NULL(input0);
50 MS_EXCEPTION_IF_NULL(input1);
51 MS_EXCEPTION_IF_NULL(input2);
52 auto abstract = pooling_grad->Infer({input0->ToAbstract(), input1->ToAbstract(), input2->ToAbstract()});
53 MS_EXCEPTION_IF_NULL(abstract);
54 EXPECT_EQ(abstract->isa<abstract::AbstractTensor>(), true);
55 auto shape_ptr = abstract->BuildShape();
56 MS_EXCEPTION_IF_NULL(shape_ptr);
57 EXPECT_EQ(shape_ptr->isa<abstract::Shape>(), true);
58 auto shape = shape_ptr->cast<abstract::ShapePtr>();
59 MS_EXCEPTION_IF_NULL(shape);
60 auto shape_vec = shape->shape();
61 EXPECT_EQ(shape_vec.size(), 2);
62 EXPECT_EQ(shape_vec[0], 3);
63 EXPECT_EQ(shape_vec[1], 3);
64 auto type = abstract->BuildType();
65 MS_EXCEPTION_IF_NULL(type);
66 EXPECT_EQ(type->isa<TensorType>(), true);
67 auto tensor_type = type->cast<TensorTypePtr>();
68 MS_EXCEPTION_IF_NULL(tensor_type);
69 auto data_type = tensor_type->element();
70 MS_EXCEPTION_IF_NULL(data_type);
71 EXPECT_EQ(data_type->type_id(), kNumberTypeFloat32);
72 }
73 } // namespace ops
74 } // namespace mindspore
75