test_imperative_gan.py 6.7 KB
Newer Older
X
Xin Pan 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
# 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 contextlib
import unittest
import numpy as np
import six
import sys

import paddle
import paddle.fluid as fluid
M
minqiyang 已提交
23
import paddle.fluid.core as core
X
Xin Pan 已提交
24 25 26
from paddle.fluid.optimizer import SGDOptimizer
from paddle.fluid.imperative.nn import Conv2D, Pool2D, FC
from test_imperative_base import new_program_scope
X
Xin Pan 已提交
27
from paddle.fluid.imperative.base import to_variable
X
Xin Pan 已提交
28 29 30


class Discriminator(fluid.imperative.Layer):
X
Xin Pan 已提交
31 32 33 34
    def __init__(self, name_scope):
        super(Discriminator, self).__init__(name_scope)
        self._fc1 = FC(self.full_name(), size=32, act='elu')
        self._fc2 = FC(self.full_name(), size=1)
X
Xin Pan 已提交
35 36 37 38 39 40 41

    def forward(self, inputs):
        x = self._fc1(inputs)
        return self._fc2(x)


class Generator(fluid.imperative.Layer):
X
Xin Pan 已提交
42 43 44 45 46
    def __init__(self, name_scope):
        super(Generator, self).__init__(name_scope)
        self._fc1 = FC(self.full_name(), size=64, act='elu')
        self._fc2 = FC(self.full_name(), size=64, act='elu')
        self._fc3 = FC(self.full_name(), size=1)
X
Xin Pan 已提交
47 48 49 50 51 52 53

    def forward(self, inputs):
        x = self._fc1(inputs)
        x = self._fc2(x)
        return self._fc3(x)


X
Xin Pan 已提交
54
class TestImperativeGAN(unittest.TestCase):
M
minqiyang 已提交
55
    def test_gan_float32(self):
X
Xin Pan 已提交
56 57 58 59 60
        seed = 90

        startup = fluid.Program()
        startup.random_seed = seed
        discriminate_p = fluid.Program()
X
Xin Pan 已提交
61 62 63 64
        generate_p = fluid.Program()
        discriminate_p.random_seed = seed
        generate_p.random_seed = seed

X
Xin Pan 已提交
65 66 67
        scope = fluid.core.Scope()
        with new_program_scope(
                main=discriminate_p, startup=startup, scope=scope):
X
Xin Pan 已提交
68 69
            discriminator = Discriminator("d")
            generator = Generator("g")
X
Xin Pan 已提交
70 71 72 73 74 75 76 77 78

            img = fluid.layers.data(
                name="img", shape=[2, 1], append_batch_size=False)
            noise = fluid.layers.data(
                name="noise", shape=[2, 2], append_batch_size=False)

            d_real = discriminator(img)
            d_loss_real = fluid.layers.reduce_mean(
                fluid.layers.sigmoid_cross_entropy_with_logits(
X
Xin Pan 已提交
79 80 81
                    x=d_real,
                    label=fluid.layers.fill_constant(
                        shape=[2, 1], dtype='float32', value=1.0)))
X
Xin Pan 已提交
82 83 84 85

            d_fake = discriminator(generator(noise))
            d_loss_fake = fluid.layers.reduce_mean(
                fluid.layers.sigmoid_cross_entropy_with_logits(
X
Xin Pan 已提交
86 87 88
                    x=d_fake,
                    label=fluid.layers.fill_constant(
                        shape=[2, 1], dtype='float32', value=0.0)))
X
Xin Pan 已提交
89 90 91 92 93 94 95

            d_loss = d_loss_real + d_loss_fake

            sgd = SGDOptimizer(learning_rate=1e-3)
            sgd.minimize(d_loss)

        with new_program_scope(main=generate_p, startup=startup, scope=scope):
X
Xin Pan 已提交
96 97
            discriminator = Discriminator("d")
            generator = Generator("g")
X
Xin Pan 已提交
98 99 100 101 102 103 104

            noise = fluid.layers.data(
                name="noise", shape=[2, 2], append_batch_size=False)

            d_fake = discriminator(generator(noise))
            g_loss = fluid.layers.reduce_mean(
                fluid.layers.sigmoid_cross_entropy_with_logits(
X
Xin Pan 已提交
105 106 107
                    x=d_fake,
                    label=fluid.layers.fill_constant(
                        shape=[2, 1], dtype='float32', value=1.0)))
X
Xin Pan 已提交
108 109 110 111

            sgd = SGDOptimizer(learning_rate=1e-3)
            sgd.minimize(g_loss)

M
minqiyang 已提交
112 113
        exe = fluid.Executor(fluid.CPUPlace() if not core.is_compiled_with_cuda(
        ) else fluid.CUDAPlace(0))
X
Xin Pan 已提交
114
        static_params = dict()
X
Xin Pan 已提交
115 116 117 118
        with fluid.scope_guard(scope):
            img = np.ones([2, 1], np.float32)
            noise = np.ones([2, 2], np.float32)
            exe.run(startup)
X
Xin Pan 已提交
119 120 121 122 123 124 125 126 127
            static_d_loss = exe.run(discriminate_p,
                                    feed={'img': img,
                                          'noise': noise},
                                    fetch_list=[d_loss])[0]
            static_g_loss = exe.run(generate_p,
                                    feed={'noise': noise},
                                    fetch_list=[g_loss])[0]

            # generate_p contains all parameters needed.
X
Xin Pan 已提交
128
            for param in generate_p.global_block().all_parameters():
X
Xin Pan 已提交
129 130 131 132
                static_params[param.name] = np.array(
                    scope.find_var(param.name).get_tensor())

        dy_params = dict()
M
minqiyang 已提交
133
        with fluid.imperative.guard():
X
Xin Pan 已提交
134 135 136
            fluid.default_startup_program().random_seed = seed
            fluid.default_main_program().random_seed = seed

X
Xin Pan 已提交
137 138
            discriminator = Discriminator("d")
            generator = Generator("g")
X
Xin Pan 已提交
139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
            sgd = SGDOptimizer(learning_rate=1e-3)

            d_real = discriminator(to_variable(np.ones([2, 1], np.float32)))
            d_loss_real = fluid.layers.reduce_mean(
                fluid.layers.sigmoid_cross_entropy_with_logits(
                    x=d_real, label=to_variable(np.ones([2, 1], np.float32))))

            d_fake = discriminator(
                generator(to_variable(np.ones([2, 2], np.float32))))
            d_loss_fake = fluid.layers.reduce_mean(
                fluid.layers.sigmoid_cross_entropy_with_logits(
                    x=d_fake, label=to_variable(np.zeros([2, 1], np.float32))))

            d_loss = d_loss_real + d_loss_fake
            d_loss._backward()
            sgd.minimize(d_loss)
X
Xin Pan 已提交
155 156
            discriminator.clear_gradients()
            generator.clear_gradients()
X
Xin Pan 已提交
157

X
Xin Pan 已提交
158 159 160 161 162 163 164 165 166 167 168
            d_fake = discriminator(
                generator(to_variable(np.ones([2, 2], np.float32))))
            g_loss = fluid.layers.reduce_mean(
                fluid.layers.sigmoid_cross_entropy_with_logits(
                    x=d_fake, label=to_variable(np.ones([2, 1], np.float32))))
            g_loss._backward()
            sgd.minimize(g_loss)
            for p in discriminator.parameters():
                dy_params[p.name] = p._numpy()
            for p in generator.parameters():
                dy_params[p.name] = p._numpy()
X
Xin Pan 已提交
169 170 171 172 173 174 175 176

            dy_g_loss = g_loss._numpy()
            dy_d_loss = d_loss._numpy()

        self.assertEqual(dy_g_loss, static_g_loss)
        self.assertEqual(dy_d_loss, static_d_loss)
        for k, v in six.iteritems(dy_params):
            self.assertTrue(np.allclose(v, static_params[k]))
X
Xin Pan 已提交
177 178 179 180


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