blob: 11a5d1eb87e4c8de40478e66d614194ee5195f59 [file] [log] [blame]
//
// Copyright © 2017 Arm Ltd. All rights reserved.
// See LICENSE file in the project root for full license information.
//
#include <boost/test/unit_test.hpp>
#include "armnnOnnxParser/IOnnxParser.hpp"
#include "ParserPrototxtFixture.hpp"
BOOST_AUTO_TEST_SUITE(OnnxParser)
struct SimpleConv2DFixture : public armnnUtils::ParserPrototxtFixture<armnnOnnxParser::IOnnxParser>
{
SimpleConv2DFixture()
{
m_Prototext = R"(
ir_version: 3
producer_name: "CNTK"
producer_version: "2.5.1"
domain: "ai.cntk"
model_version: 1
graph {
name: "CNTKGraph"
input {
name: "Input"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 1
}
dim {
dim_value: 1
}
dim {
dim_value: 3
}
dim {
dim_value: 3
}
}
}
}
}
input {
name: "Weight"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 1
}
dim {
dim_value: 1
}
dim {
dim_value: 3
}
dim {
dim_value: 3
}
}
}
}
}
initializer {
dims: 1
dims: 1
dims: 3
dims: 3
data_type: FLOAT
float_data: 2
float_data: 1
float_data: 0
float_data: 6
float_data: 2
float_data: 1
float_data: 4
float_data: 1
float_data: 2
name: "Weight"
}
node {
input: "Input"
input: "Weight"
output: "Output"
name: "Convolution"
op_type: "Conv"
attribute {
name: "kernel_shape"
ints: 3
ints: 3
type: INTS
}
attribute {
name: "strides"
ints: 1
ints: 1
type: INTS
}
attribute {
name: "auto_pad"
s: "VALID"
type: STRING
}
attribute {
name: "group"
i: 1
type: INT
}
attribute {
name: "dilations"
ints: 1
ints: 1
type: INTS
}
doc_string: ""
domain: ""
}
output {
name: "Output"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 1
}
dim {
dim_value: 1
}
dim {
dim_value: 1
}
dim {
dim_value: 1
}
}
}
}
}
}
opset_import {
version: 7
})";
Setup();
}
};
struct Conv2DWithBiasesFixture : public armnnUtils::ParserPrototxtFixture<armnnOnnxParser::IOnnxParser>
{
Conv2DWithBiasesFixture() {
m_Prototext = R"(
ir_version: 3
producer_name: "CNTK"
producer_version: "2.5.1"
domain: "ai.cntk"
model_version: 1
graph {
name: "CNTKGraph"
input {
name: "Input"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 1
}
dim {
dim_value: 1
}
dim {
dim_value: 2
}
dim {
dim_value: 2
}
}
}
}
}
input {
name: "Weight"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 1
}
dim {
dim_value: 1
}
dim {
dim_value: 2
}
dim {
dim_value: 2
}
}
}
}
}
initializer {
dims: 1
dims: 1
dims: 2
dims: 2
data_type: FLOAT
float_data: 2
float_data: 1
float_data: 0
float_data: 6
name: "Weight"
}
input {
name: "Bias"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 4
}
}
}
}
}
initializer {
dims: 4
data_type: FLOAT
float_data: 10
float_data: 0
float_data: 0
float_data: 0
name: "Bias"
}
node {
input: "Input"
input: "Weight"
input: "Bias"
output: "Output"
name: "Convolution"
op_type: "Conv"
attribute {
name: "kernel_shape"
ints: 2
ints: 2
type: INTS
}
attribute {
name: "strides"
ints: 1
ints: 1
type: INTS
}
attribute {
name: "auto_pad"
s: "SAME_UPPER"
type: STRING
}
attribute {
name: "group"
i: 1
type: INT
}
attribute {
name: "dilations"
ints: 1
ints: 1
type: INTS
}
doc_string: ""
domain: ""
}
output {
name: "Output"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 1
}
dim {
dim_value: 1
}
dim {
dim_value: 2
}
dim {
dim_value: 2
}
}
}
}
}
}
opset_import {
version: 7
})";
Setup();
}
};
struct Conv2DDimReducingFixture : public armnnUtils::ParserPrototxtFixture<armnnOnnxParser::IOnnxParser>
{
Conv2DDimReducingFixture() {
m_Prototext = R"(
ir_version: 3
producer_name: "CNTK"
producer_version: "2.5.1"
domain: "ai.cntk"
model_version: 1
graph {
name: "CNTKGraph"
input {
name: "Input"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 1
}
dim {
dim_value: 3
}
dim {
dim_value: 2
}
dim {
dim_value: 2
}
}
}
}
}
input {
name: "Weight"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 2
}
dim {
dim_value: 3
}
dim {
dim_value: 1
}
dim {
dim_value: 1
}
}
}
}
}
initializer {
dims: 2
dims: 3
dims: 1
dims: 1
data_type: FLOAT
float_data: -1
float_data: 2
float_data: 0
float_data: 1
float_data: 0
float_data: 0
name: "Weight"
}
node {
input: "Input"
input: "Weight"
output: "Output"
name: "Convolution"
op_type: "Conv"
attribute {
name: "kernel_shape"
ints: 1
ints: 1
type: INTS
}
attribute {
name: "strides"
ints: 1
ints: 1
type: INTS
}
attribute {
name: "group"
i: 1
type: INT
}
attribute {
name: "dilations"
ints: 1
ints: 1
type: INTS
}
doc_string: ""
domain: ""
}
output {
name: "Output"
type {
tensor_type {
elem_type: FLOAT
shape {
dim {
dim_value: 1
}
dim {
dim_value: 2
}
dim {
dim_value: 2
}
dim {
dim_value: 2
}
}
}
}
}
}
opset_import {
version: 7
})";
Setup();
}
};
BOOST_FIXTURE_TEST_CASE(ValidConvTest, SimpleConv2DFixture)
{
RunTest<4>({{"Input", {1.0, 2.0, 3.0,
4.0, 5.0, 6.0,
7.0, 8.0, 9.0}}},
{{"Output", {1.0 * 2 + 2.0 * 1 + 3.0 * 0 +
4.0 * 6 + 5.0 * 2 + 6.0 * 1 +
7.0 * 4 + 8.0 * 1 + 9.0 * 2}}});
}
BOOST_FIXTURE_TEST_CASE(ValidConvWithBiasTest, Conv2DWithBiasesFixture)
{
RunTest<4>({{"Input", {1.0, 2.0,
3.0, 4.0}}},
{{"Output", {1.0 * 2 + 2.0 * 1 + 3.0 * 0 + 4 * 6 + 10,
2.0 * 2 + 0 * 1 + 4.0 * 0 + 0 * 6 + 10,
3.0 * 2 + 4.0 * 1 + 0 * 0 + 0 * 6 + 10,
4.0 * 2 + 0 * 1 + 0 * 0 + 0 * 6 + 10}}});
}
BOOST_FIXTURE_TEST_CASE(ValidConvDimReducTest, Conv2DDimReducingFixture)
{
RunTest<4>({{"Input", {1.0, 2.0, 3.0, 4.0, -1, -2, 3, 4, 1 , 1, 1, 1 }}},
{{"Output", {-1 * 1 + 2 * -1, -1 * 2 + 2 * -2,
-1 * 3 + 2 * 3, -1 * 4 + 2 * 4,
1, 2, 3, 4}}});
}
BOOST_AUTO_TEST_SUITE_END()