mindspore/tests/st/ops/cpu/test_conv_op.py

390 lines
17 KiB
Python

# Copyright 2019 Huawei Technologies Co., Ltd
#
# 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 numpy as np
import pytest
import mindspore.context as context
import mindspore.nn as nn
from mindspore import Tensor
from mindspore.common.initializer import initializer
from mindspore.common.parameter import Parameter
from mindspore.common.parameter import ParameterTuple
from mindspore.ops import operations as P
from mindspore.ops import composite as C
context.set_context(mode=context.GRAPH_MODE, device_target='CPU')
class NetConv2d(nn.Cell):
def __init__(self):
super(NetConv2d, self).__init__()
out_channel = 2
kernel_size = 1
self.conv = P.Conv2D(out_channel,
kernel_size,
mode=1,
pad_mode="valid",
pad=0,
stride=1,
dilation=1,
group=1)
self.w = Parameter(initializer(
Tensor(np.arange(2 * 3 * 1 * 1).reshape(2, 3, 1, 1).astype(np.float32)), [2, 3, 1, 1]), name='w')
self.x = Parameter(initializer(
Tensor(np.arange(1 * 3 * 3 * 3).reshape(1, 3, 3, 3).astype(np.float32)), [1, 3, 3, 3]), name='x')
def construct(self):
return self.conv(self.x, self.w)
@pytest.mark.level0
@pytest.mark.platform_x86_cpu
@pytest.mark.env_onecard
def test_conv2d():
conv2d = NetConv2d()
output = conv2d()
expect = np.array([[[[45, 48, 51],
[54, 57, 60],
[63, 66, 69]],
[[126, 138, 150],
[162, 174, 186],
[198, 210, 222]]]]).astype(np.float32)
assert (output.asnumpy() == expect).all()
class NetConv(nn.Cell):
def __init__(self, weight, x):
super(NetConv, self).__init__()
self.conv = nn.Conv2d(in_channels=3,
out_channels=3,
kernel_size=(5, 3),
stride=2,
pad_mode='same',
padding=(0, 0, 0, 0),
dilation=(1, 1),
group=1,
has_bias=False,
weight_init=Tensor(weight)
)
self.x = Parameter(initializer(Tensor(x), [1, 3, 4, 2]), name="x")
def construct(self):
return self.conv(self.x)
@pytest.mark.level0
@pytest.mark.platform_x86_cpu
@pytest.mark.env_onecard
def test_conv():
weight = np.array([[[[0.38968208, 0.14398979, 0.7962463],
[-2.1836321, -0.63823014, -0.50588065],
[0.6660469, 0.64673275, -0.13160042],
[1.3683757, 1.4005762, -0.37235805],
[-0.22638111, 0.45427424, -0.10293389]],
[[1.4985064, -0.29318333, -0.92694616],
[1.539068, 0.8937254, -1.2598171],
[0.9658142, -0.63945454, -0.23185322],
[1.363089, -0.41694695, -2.2750475],
[-0.4865508, -1.6938025, 0.609849]],
[[1.1844803, 0.99874926, -1.9475793],
[0.4987858, 0.5307887, -0.04226681],
[0.4529779, -1.1960793, 0.9456575],
[3.133675, 0.2309789, -0.29201075],
[-0.59632736, -0.0789804, -0.69486314]]],
[[[-0.5606142, 0.6420862, 0.2478745],
[0.02717604, 1.5483379, -0.9373383],
[-1.1017276, -0.259478, 1.0311872],
[1.8387799, 0.16468556, 0.33392152],
[-1.8781787, 1.0158662, 1.6527579]],
[[0.45696944, -0.5652523, -1.5618048],
[-0.30304828, 0.1331878, -0.36955845],
[0.91655576, 0.66612357, 0.3068175],
[-0.45732066, 0.8923335, 1.0542952],
[-0.73519516, 1.0518405, -1.0273266]],
[[-0.79712886, -0.26814285, 0.12779616],
[1.0367643, -1.6180774, 0.42999932],
[-0.81818223, -0.81502074, 0.882194],
[0.53640485, 0.4178927, 1.6037121],
[0.9256354, -1.1006796, 0.16614541]]],
[[[-1.5216796, -1.2473261, 0.6549515],
[0.63627815, 0.7221449, 0.02977821],
[-0.61331123, -0.49451825, 0.33852202],
[1.4510741, -1.3818305, -0.791747],
[0.6989747, 0.49558765, 1.0813237]],
[[-0.03969796, 0.71586496, 0.8326594],
[-0.15443641, 1.0389746, -0.59301984],
[0.7197836, 0.03257621, 1.8398637],
[0.6111736, -0.16166899, -2.4869773],
[1.3066711, -1.8003578, 0.17412892]],
[[-0.31470737, -0.5938182, -1.1311078],
[-0.99081016, 0.4005125, 0.44154453],
[1.0876914, -2.5958562, -0.5914863],
[1.3759689, -0.7741513, 0.19928917],
[1.6792973, 2.2744863, -0.04308867]]]]).astype(np.float32)
x = np.array([[[[-1.4311737, 1.015344],
[0.04431088, -2.2886624],
[1.4832113, 1.240908],
[0.67040104, 0.15266363]],
[[0.44226435, 1.1461105],
[1.194218, 1.5547837],
[0.23152256, 1.5911953],
[0.11206784, 0.17978816]],
[[-0.57803905, 0.8039611],
[0.0823025, -0.6134477],
[-1.4171146, 1.6269946],
[0.48878875, 0.9117505]]]]).astype(np.float32)
conv2d = NetConv(weight, x)
output = conv2d()
expected = np.array([[[[2.3498724],
[-1.9199573]],
[[5.376562],
[-5.425745]],
[[5.9105043],
[7.469034]]]]).astype(np.float32)
loss = np.abs(expected - output.asnumpy())
error = 1e-4 * np.ones(loss.shape)
assert (loss < error).all()
class NetConv3d(nn.Cell):
def __init__(self, mode, pad_mode, pad):
super(NetConv3d, self).__init__()
out_channel = 4
kernel_size = 2
self.conv = P.Conv3D(out_channel,
kernel_size,
mode=mode,
pad_mode=pad_mode,
pad=pad,
stride=1,
dilation=1,
group=1)
def construct(self, x, w):
return self.conv(x, w)
@pytest.mark.level0
@pytest.mark.platform_x86_cpu
@pytest.mark.env_onecard
def test_conv3d():
x = Tensor(np.arange(1 * 3 * 3 * 3 * 3).reshape(1, 3, 3, 3, 3).astype(np.float32))
w = Tensor(np.arange(4 * 3 * 2 * 2 * 2).reshape(4, 3, 2, 2, 2).astype(np.float32))
expect = np.array([[[[[12960., 13236.],
[13788., 14064.]],
[[15444., 15720.],
[16272., 16548.]]],
[[[32256., 33108.],
[34812., 35664.]],
[[39924., 40776.],
[42480., 43332.]]],
[[[51552., 52980.],
[55836., 57264.]],
[[64404., 65832.],
[68688., 70116.]]],
[[[70848., 72852.],
[76860., 78864.]],
[[88884., 90888.],
[94896., 96900.]]]]]).astype(np.float32)
mode = 1
pad_mode = "valid"
pad = 0
net = NetConv3d(mode, pad_mode, pad)
output = net(x, w)
assert (output.asnumpy() == expect).all()
@pytest.mark.level0
@pytest.mark.platform_x86_cpu
@pytest.mark.env_onecard
def test_conv3d_2():
x = Tensor(np.arange(1 * 3 * 3 * 3 * 3).reshape(1, 3, 3, 3, 3).astype(np.float32))
w = Tensor(np.arange(4 * 3 * 2 * 2 * 2).reshape(4, 3, 2, 2, 2).astype(np.float32))
expect = np.array([[[[[1647, 3258, 3345, 1650],
[3267, 6447, 6609, 3252],
[3519, 6933, 7095, 3486],
[1719, 3378, 3453, 1692]],
[[3375, 6639, 6789, 3330],
[6606, 12960, 13236, 6474],
[7038, 13788, 14064, 6870],
[3393, 6627, 6753, 3288]],
[[4077, 7989, 8139, 3978],
[7902, 15444, 15720, 7662],
[8334, 16272, 16548, 8058],
[3987, 7761, 7887, 3828]],
[[1917, 3732, 3795, 1842],
[3663, 7107, 7221, 3492],
[3843, 7449, 7563, 3654],
[1809, 3492, 3543, 1704]]],
[[[3591, 7218, 7449, 3738],
[7371, 14799, 15249, 7644],
[8055, 16149, 16599, 8310],
[4095, 8202, 8421, 4212]],
[[7911, 15855, 16293, 8154],
[16110, 32256, 33108, 16554],
[17406, 34812, 35664, 17814],
[8793, 17571, 17985, 8976]],
[[9909, 19797, 20235, 10098],
[19998, 39924, 40776, 20334],
[21294, 42480, 43332, 21594],
[10683, 21297, 21711, 10812]],
[[5157, 10284, 10491, 5226],
[10359, 20643, 21045, 10476],
[10971, 21849, 22251, 11070],
[5481, 10908, 11103, 5520]]],
[[[5535, 11178, 11553, 5826],
[11475, 23151, 23889, 12036],
[12591, 25365, 26103, 13134],
[6471, 13026, 13389, 6732]],
[[12447, 25071, 25797, 12978],
[25614, 51552, 52980, 26634],
[27774, 55836, 57264, 28758],
[14193, 28515, 29217, 14664]],
[[15741, 31605, 32331, 16218],
[32094, 64404, 65832, 33006],
[34254, 68688, 70116, 35130],
[17379, 34833, 35535, 17796]],
[[8397, 16836, 17187, 8610],
[17055, 34179, 34869, 17460],
[18099, 36249, 36939, 18486],
[9153, 18324, 18663, 9336]]],
[[[7479, 15138, 15657, 7914],
[15579, 31503, 32529, 16428],
[17127, 34581, 35607, 17958],
[8847, 17850, 18357, 9252]],
[[16983, 34287, 35301, 17802],
[35118, 70848, 72852, 36714],
[38142, 76860, 78864, 39702],
[19593, 39459, 40449, 20352]],
[[21573, 43413, 44427, 22338],
[44190, 88884, 90888, 45678],
[47214, 94896, 96900, 48666],
[24075, 48369, 49359, 24780]],
[[11637, 23388, 23883, 11994],
[23751, 47715, 48693, 24444],
[25227, 50649, 51627, 25902],
[12825, 25740, 26223, 13152]]]]]).astype(np.float32)
mode = 1
pad_mode = "pad"
pad = (1, 1, 1, 1, 1, 1)
net = NetConv3d(mode, pad_mode, pad)
output = net(x, w)
assert (output.asnumpy() == expect).all()
class Conv3dNet(nn.Cell):
def __init__(self, in_channels, out_channels, kernel_size, pad_mode='pad', padding=0, stride=1, dilation=1,
has_bias=False, weight_init='normal'):
super(Conv3dNet, self).__init__()
self.cv1 = nn.Conv3d(in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
pad_mode=pad_mode,
padding=padding,
stride=stride,
dilation=dilation,
group=1,
has_bias=has_bias,
weight_init=weight_init,
data_format='NCDHW')
def construct(self, x):
x = self.cv1(x)
return x
class GradNet(nn.Cell):
def __init__(self, network):
super(GradNet, self).__init__()
self.grad = C.GradOperation(get_all=True, sens_param=True, get_by_list=True)
self.network = network
self.params = ParameterTuple(network.trainable_params())
def construct(self, x, dy):
grad_op = self.grad(self.network, self.params)
output = grad_op(x, dy)
return output
@pytest.mark.level0
@pytest.mark.platform_x86_cpu
@pytest.mark.env_onecard
def test_conv3d_with_grad():
"""
Feature: test conv3d op.
Description: including forward and backward.
Expectation: expect correct forward and backward result.
"""
np_type = np.float32
x = Tensor(np.array([[[[[1.6924546, 0.05080776, -0.6369957],
[0.19091548, 2.1002553, 0.12015896],
[0.6172031, 0.30017033, -0.35224986]],
[[-1.1425182, -0.34934273, -0.20889424],
[0.5866232, 0.8389834, 0.9311021],
[0.2855873, 0.8851412, -0.7543979]],
[[1.2528682, 0.5129298, -0.29809284],
[0.48851815, -0.07557172, 1.1316293],
[1.5198169, 2.1855755, -1.3964963]]]]]).astype(np_type))
dy = Tensor(np.array([[[[[-1.4441139, -0.5044659],
[0.16003707, 0.8761689]],
[[0.31563494, -2.0222013],
[-0.30620402, 0.8279746]]],
[[[0.23009473, 0.7620112],
[-0.22232814, -0.20075807]],
[[0.18656139, 0.41005164],
[0.19829972, 0.11900865]]]]]).astype(np_type))
w = Tensor(np.array([[[[[-0.9358, -0.2679],
[0.5304, -0.6917]],
[[-0.3968, -0.6872],
[-0.8452, -0.6712]]]],
[[[[-0.0127, -1.1173],
[0.2344, 1.6598]],
[[0.7420, -0.1918],
[-0.8876, -0.7472]]]]]).astype(np_type))
net = Conv3dNet(in_channels=x.shape[1], out_channels=2, kernel_size=(2, 2, 2), weight_init=w)
actual_out = net(x)
expect_out = np.array([[[[[-3.3144155, 0.10207337],
[-2.266387, -2.8092794]],
[[-0.31821766, -0.51052636],
[-4.127921, -1.700856]]],
[[[1.524173, -0.2567379],
[-2.346652, -0.4532562]],
[[2.3889866, 1.6392273],
[-2.0138235, -3.2652235]]]]]).astype(np_type)
assert np.allclose(actual_out.asnumpy(), expect_out)
grad_net = GradNet(net)
actual_grads = grad_net(x, dy)
expect_dx = np.array([[[[[1.3484796, 0.5921949, -0.71624875],
[-0.8589629, 0.68001556, 1.6033065],
[0.03276994, -0.06205562, -0.9392643]],
[[0.44601417, 3.308012, 0.2841122],
[1.2830329, -1.8175733, 0.93020254],
[-0.05385402, 0.5043542, -0.81325763]],
[[0.01318461, 0.85398096, 1.3110089],
[-0.16372639, 0.9261035, 0.45910096],
[0.0827928, -0.7480816, -0.6446598]]]]]).astype(np_type)
expect_dw = np.array([[[[[0.26185727, 1.515559],
[-1.8394437, -5.867935]],
[[1.8011744, 3.2847447],
[1.2020903, -6.338352]]]],
[[[[-0.17617291, -0.8384279],
[2.0623026, 1.2028661]],
[[-0.29600215, -0.5198703],
[1.1547322, 1.5743471]]]]]).astype(np_type)
assert np.allclose(actual_grads[0][0].asnumpy(), expect_dx)
assert np.allclose(actual_grads[1][0].asnumpy(), expect_dw)