test_deconv2d.py

import unittest
import numpy as np
import dezero.layers as L
import dezero.functions as F
from dezero.utils import gradient_check, array_allclose
import chainer.functions as CF


class TestDeconv2d(unittest.TestCase):

    def test_forward1(self):
        n, c_i, c_o = 10, 1, 3
        h_i, w_i = 5, 10
        h_k, w_k = 10, 10
        h_p, w_p = 5, 5
        s_y, s_x = 5, 5
        x = np.random.uniform(0, 1, (n, c_i, h_i, w_i)).astype(np.float32)
        W = np.random.uniform(0, 1, (c_i, c_o, h_k, w_k)).astype(np.float32)
        b = np.random.uniform(0, 1, c_o).astype(np.float32)

        expected = CF.deconvolution_2d(x, W, b, stride=(s_y, s_x),
                                       pad=(h_p, w_p))
        y = F.deconv2d(x, W, b, stride=(s_y, s_x), pad=(h_p, w_p))
        self.assertTrue(array_allclose(expected.data, y.data))

    def test_forward2(self):
        n, c_i, c_o = 10, 1, 3
        h_i, w_i = 5, 10
        h_k, w_k = 10, 10
        h_p, w_p = 5, 5
        s_y, s_x = 5, 5
        x = np.random.uniform(0, 1, (n, c_i, h_i, w_i)).astype(np.float32)
        W = np.random.uniform(0, 1, (c_i, c_o, h_k, w_k)).astype(np.float32)
        b = None
        expected = CF.deconvolution_2d(x, W, b, stride=(s_y, s_x),
                                       pad=(h_p, w_p))
        y = F.deconv2d(x, W, b, stride=(s_y, s_x), pad=(h_p, w_p))
        self.assertTrue(array_allclose(expected.data, y.data))

    def test_backward1(self):
        n, c_i, c_o = 10, 1, 3
        h_i, w_i = 5, 10
        h_k, w_k = 10, 10
        h_p, w_p = 5, 5
        s_y, s_x = 5, 5
        x = np.random.uniform(0, 1, (n, c_i, h_i, w_i))
        W = np.random.uniform(0, 1, (c_i, c_o, h_k, w_k))
        b = None  # np.random.uniform(0, 1, c_o).astype(np.float32)
        f = lambda x: F.deconv2d(x, W, b, stride=(s_y, s_x), pad=(h_p, w_p))
        self.assertTrue(gradient_check(f, x))

    def test_backward2(self):
        n, c_i, c_o = 10, 1, 3
        h_i, w_i = 5, 10
        h_k, w_k = 10, 10
        h_p, w_p = 5, 5
        s_y, s_x = 5, 5
        x = np.random.uniform(0, 1, (n, c_i, h_i, w_i))
        W = np.random.uniform(0, 1, (c_i, c_o, h_k, w_k))
        b = np.random.uniform(0, 1, c_o)
        f = lambda W: F.deconv2d(x, W, b, stride=(s_y, s_x), pad=(h_p, w_p))
        self.assertTrue(gradient_check(f, W))

    def test_backward3(self):
        n, c_i, c_o = 10, 1, 3
        h_i, w_i = 5, 10
        h_k, w_k = 10, 10
        h_p, w_p = 5, 5
        s_y, s_x = 5, 5
        x = np.random.uniform(0, 1, (n, c_i, h_i, w_i))
        W = np.random.uniform(0, 1, (c_i, c_o, h_k, w_k))
        b = np.random.uniform(0, 1, c_o)
        f = lambda b: F.deconv2d(x, W, b, stride=(s_y, s_x), pad=(h_p, w_p))
        self.assertTrue(gradient_check(f, b))