• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1# Copyright 2021 Huawei Technologies Co., Ltd
2#
3# Licensed under the Apache License, Version 2.0 (the "License");
4# you may not use this file except in compliance with the License.
5# You may obtain a copy of the License at
6#
7# http://www.apache.org/licenses/LICENSE-2.0
8#
9# Unless required by applicable law or agreed to in writing, software
10# distributed under the License is distributed on an "AS IS" BASIS,
11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12# See the License for the specific language governing permissions and
13# limitations under the License.
14# ============================================================================
15
16import numpy as np
17import pytest
18
19import mindspore.context as context
20import mindspore.nn as nn
21from mindspore import Tensor
22from mindspore.ops import operations as P
23from mindspore.common import dtype as mstype
24from mindspore.ops.composite import GradOperation
25
26
27class Net(nn.Cell):
28    def __init__(self):
29        super(Net, self).__init__()
30        self.loss = P.CTCLoss()
31        self.div = P.RealDiv()
32        self.mean = P.ReduceMean()
33
34    def construct(self, probs, label, input_length, indices):
35        x, _ = self.loss(probs, indices, label, input_length)
36        x = self.mean(x)
37        return x
38
39
40class GradData(nn.Cell):
41    def __init__(self, network):
42        super(GradData, self).__init__()
43        self.grad = GradOperation(get_all=True, sens_param=False)
44        self.network = network
45
46    def construct(self, probs, indices, labels, input_lengths):
47        return self.grad(self.network)(probs, indices, labels, input_lengths)
48
49
50@pytest.mark.level0
51@pytest.mark.platform_x86_cpu
52@pytest.mark.env_onecard
53def test_ctcloss():
54    probs = Tensor([[[-4.4131, -4.6093, -3.4333, -3.9268, -2.8917, -3.4093, -4.2243, -1.1379, -7.1046, -0.6902],
55                     [-2.5109, -3.3397, -4.9384, -1.2723, -1.1443, -2.4683, -2.6768, -4.1282, -2.7062, -3.1906],
56                     [-2.5092, -1.6392, -2.0864, -4.0059, -1.5610, -2.3223, -2.4816, -2.9922, -3.1412, -2.3311]],
57
58                    [[-2.1243, -3.5773, -3.1108, -4.4253, -2.7080, -1.9653, -2.0499, -2.4418, -1.8620, -1.5229],
59                     [-2.2479, -3.5128, -1.4189, -2.8701, -1.8562, -2.2752, -2.7019, -2.1865, -2.5634, -2.9869],
60                     [-3.2144, -1.3986, -3.1083, -3.9634, -3.5131, -3.2317, -2.6200, -1.7938, -1.8159, -1.7255]],
61
62                    [[-3.1301, -2.1649, -0.9286, -2.9452, -2.5992, -2.0263, -2.9201, -3.2155, -2.8302, -3.3636],
63                     [-1.4661, -3.6311, -2.4781, -4.6180, -2.7308, -1.7019, -1.5570, -2.6012, -4.0788, -2.3073],
64                     [-2.6833, -1.5033, -3.6922, -2.6360, -2.6974, -2.6847, -2.7579, -2.1396, -1.4093, -2.9630]],
65
66                    [[-2.0094, -2.3024, -3.3673, -1.0220, -2.8326, -2.2613, -3.0535, -2.9879, -3.7015, -2.4510],
67                     [-1.9071, -3.2603, -2.3229, -2.0572, -4.3450, -2.1284, -2.6306, -1.3824, -2.9815, -2.5061],
68                     [-2.7931, -3.7631, -3.2440, -4.3887, -1.0271, -3.8851, -1.2418, -4.5123, -2.2993, -2.4607]],
69
70                    [[-1.5763, -2.7539, -3.6941, -3.8166, -1.2599, -2.6903, -2.5826, -4.8208, -2.9562, -1.6321],
71                     [-3.3031, -3.0087, -1.9982, -1.9081, -3.8731, -2.8764, -2.2485, -2.3808, -1.4283, -2.1625],
72                     [-2.4516, -3.2394, -4.2053, -4.3541, -2.5229, -4.0717, -1.4894, -2.3151, -1.1098, -2.3465]]],
73                   dtype=mstype.float32)
74    labels = Tensor([3, 4, 6, 4, 7, 1, 4, 6, 6, 8], dtype=mstype.int32)
75    indices = [[0, 0], [0, 1], [0, 2], [1, 0], [1, 1], [1, 2], [2, 0], [2, 1], [2, 2], [2, 3]]
76    indices = Tensor(indices, dtype=mstype.int64)
77    input_lengths = Tensor([5, 5, 5], dtype=mstype.int32)
78
79    context.set_context(mode=context.GRAPH_MODE, device_target="CPU")
80    net = Net()
81    ctc_loss = net(probs, labels, input_lengths, indices)
82    expect_loss = [9.083767]
83    assert np.allclose(ctc_loss.asnumpy(), expect_loss)
84
85    grad = GradData(net)(probs, labels, input_lengths, indices)
86    grad = P.ReduceMean()(grad[0])
87    expect_grad = [-5.9604646e-09]
88    assert np.allclose(grad.asnumpy(), expect_grad, atol=1e-5)
89