mindspore/tests/ut/python/nn/test_nn_padding.py

# Copyright 2020 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.
# ============================================================================
""" test nn pad """
import numpy as np

import mindspore.nn as nn
from mindspore import Tensor
from mindspore.nn import ConstantPad1d, ConstantPad2d, ConstantPad3d, ZeroPad2d
from mindspore.ops.composite import GradOperation


class ConstantPad1dNet(nn.Cell):
    def __init__(self, padding, value):
        super(ConstantPad1dNet, self).__init__()
        self.pad = ConstantPad1d(padding, value)
        self.value = value

    def construct(self, x):
        return self.pad(x)


class ConstantPad2dNet(nn.Cell):
    def __init__(self, padding, value):
        super(ConstantPad2dNet, self).__init__()
        self.pad = ConstantPad2d(padding, value)
        self.value = value

    def construct(self, x):
        return self.pad(x)


class ConstantPad3dNet(nn.Cell):
    def __init__(self, padding, value):
        super(ConstantPad3dNet, self).__init__()
        self.pad = ConstantPad3d(padding, value)
        self.value = value

    def construct(self, x):
        return self.pad(x)


class ZeroPad2dNet(nn.Cell):
    def __init__(self, padding):
        super(ZeroPad2dNet, self).__init__()
        self.pad = ZeroPad2d(padding)

    def construct(self, x):
        return self.pad(x)


class Grad(nn.Cell):
    def __init__(self, network):
        super(Grad, self).__init__()
        self.grad = GradOperation(get_all=True, sens_param=True)
        self.network = network

    def construct(self, x, grads):
        return self.grad(self.network)(x, grads)


def test_constant_pad_1d_infer():
    """
    Feature: ConstantPad1d
    Description: Infer process of ConstantPad1d with three type parameters.
    Expectation: success
    """
    x = np.ones(shape=(1, 2, 3, 4)).astype(np.float32)
    print("=================case 1====================")
    padding = (0, 1)
    value = 0.5
    net = ConstantPad1dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)

    print("=================case 2====================")
    padding = 1
    value = 0.5
    net = ConstantPad1dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)

    print("=================case 3====================")
    padding = (-1, 0)
    value = 0.5
    net = ConstantPad1dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)


def test_constant_pad_1d_train():
    """
    Feature: ConstantPad1d
    Description: Train process of ConstantPad1d with three type parameters.
    Expectation: success
    """
    x = np.ones(shape=(1, 2, 3, 4)).astype(np.float32)
    print("=================case 1====================")
    padding = (0, 1)
    value = 0.5
    grads = np.random.random(size=(1, 2, 3, 5)).astype(np.float32)
    grad = Grad(ConstantPad1dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)

    print("=================case 2====================")
    padding = 1
    value = 0.5
    grads = np.random.random(size=(1, 2, 3, 6)).astype(np.float32)
    grad = Grad(ConstantPad1dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)

    print("=================case 3====================")
    padding = (-1, 0)
    value = 0.5
    grads = np.random.random(size=(1, 2, 3, 3)).astype(np.float32)
    grad = Grad(ConstantPad1dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)


def test_constant_pad_2d_infer():
    """
    Feature: ConstantPad2d
    Description: Infer process of ConstantPad2d with three type parameters.
    Expectation: success
    """
    x = np.ones(shape=(1, 2, 3, 4)).astype(np.float32)
    print("=================case 1====================")
    padding = (0, 1)
    value = 0.5
    net = ConstantPad2dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)

    print("=================case 2====================")
    padding = 1
    value = 0.5
    net = ConstantPad2dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)

    print("=================case 3====================")
    padding = (-1, 1, 0, 1)
    value = 0.5
    net = ConstantPad2dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)


def test_constant_pad_2d_train():
    """
    Feature: ConstantPad3d
    Description: Train process of ConstantPad2d with three type parameters.
    Expectation: success
    """
    x = np.ones(shape=(1, 2, 3, 4)).astype(np.float32)
    print("=================case 1====================")
    padding = (0, 1)
    value = 0.5
    grads = np.random.random(size=(1, 2, 3, 5)).astype(np.float32)
    grad = Grad(ConstantPad2dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)

    print("=================case 2====================")
    padding = 1
    value = 0.5
    grads = np.random.random(size=(1, 2, 5, 6)).astype(np.float32)
    grad = Grad(ConstantPad2dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)

    print("=================case 3====================")
    padding = (-1, 1, 0, 1)
    value = 0.5
    grads = np.random.random(size=(1, 2, 4, 4)).astype(np.float32)
    grad = Grad(ConstantPad2dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)


def test_constant_pad_3d_infer():
    """
    Feature: ConstantPad3d
    Description: Infer process of ConstantPad3d with three type parameters.
    Expectation: success
    """
    x = np.ones(shape=(1, 2, 3, 4)).astype(np.float32)
    print("=================case 1====================")
    padding = (0, 1)
    value = 0.5
    net = ConstantPad3dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)

    print("=================case 2====================")
    padding = 1
    value = 0.5
    net = ConstantPad3dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)

    print("=================case 3====================")
    padding = (-1, 1, 0, 1, 1, 0)
    value = 0.5
    net = ConstantPad3dNet(padding, value)
    output = net(Tensor(x))
    print(output)
    print(output.shape)


def test_constant_pad_3d_train():
    """
    Feature: ConstantPad3d
    Description: Train process of ConstantPad3d with three type parameters.
    Expectation: success
    """
    x = np.ones(shape=(1, 2, 3, 4)).astype(np.float32)
    print("=================case 1====================")
    padding = (0, 1)
    value = 0.5
    grads = np.random.random(size=(1, 2, 3, 5)).astype(np.float32)
    grad = Grad(ConstantPad3dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)

    print("=================case 2====================")
    padding = 1
    value = 0.5
    grads = np.random.random(size=(1, 4, 5, 6)).astype(np.float32)
    grad = Grad(ConstantPad3dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)

    print("=================case 3====================")
    padding = (-1, 1, 0, 1, 1, 0)
    value = 0.5
    grads = np.random.random(size=(1, 3, 4, 4)).astype(np.float32)
    grad = Grad(ConstantPad3dNet(padding, value))
    output = grad(Tensor(x), Tensor(grads))
    print(output)


def test_zero_pad_2d_infer():
    """
    Feature: ZeroPad2d
    Description: Infer process of ZeroPad2d with three type parameters.
    Expectation: success
    """
    x = np.ones(shape=(1, 2, 3, 4)).astype(np.float32)
    print("=================case 1====================")
    padding = (0, 1)
    net = ZeroPad2dNet(padding)
    output = net(Tensor(x))
    print(output)
    print(output.shape)

    print("=================case 2====================")
    padding = 1
    net = ZeroPad2dNet(padding)
    output = net(Tensor(x))
    print(output)
    print(output.shape)

    print("=================case 3====================")
    padding = (-1, 1, 0, 1)
    net = ZeroPad2dNet(padding)
    output = net(Tensor(x))
    print(output)
    print(output.shape)


def test_zero_pad_2d_train():
    """
    Feature: ZeroPad2d
    Description: Train process of ZeroPad2d with three type parameters.
    Expectation: success
    """

    x = np.ones(shape=(1, 2, 3, 4)).astype(np.float32)
    print("=================case 1====================")
    padding = (0, 1)
    grads = np.random.random(size=(1, 2, 3, 5)).astype(np.float32)
    grad = Grad(ZeroPad2dNet(padding))
    output = grad(Tensor(x), Tensor(grads))
    print(output)

    print("=================case 2====================")
    padding = 1
    grads = np.random.random(size=(1, 2, 5, 6)).astype(np.float32)
    grad = Grad(ZeroPad2dNet(padding))
    output = grad(Tensor(x), Tensor(grads))
    print(output)

    print("=================case 3====================")
    padding = (-1, 1, 0, 1)
    grads = np.random.random(size=(1, 2, 4, 4)).astype(np.float32)
    grad = Grad(ZeroPad2dNet(padding))
    output = grad(Tensor(x), Tensor(grads))
    print(output)