• 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# ============================================================================
15import numpy as np
16import pytest
17
18import mindspore as ms
19from mindspore import context, Tensor, Parameter
20from mindspore.nn import Cell, Momentum
21from mindspore.ops import operations as P
22from mindspore.train import Model
23from tests.dataset_mock import MindData
24
25
26class Dataset(MindData):
27    def __init__(self, predict, label, length=3):
28        super(Dataset, self).__init__(size=length)
29        self.predict = predict
30        self.label = label
31        self.index = 0
32        self.length = length
33
34    def __iter__(self):
35        return self
36
37    def __next__(self):
38        if self.index >= self.length:
39            raise StopIteration
40        self.index += 1
41        return self.predict, self.label
42
43    def reset(self):
44        self.index = 0
45
46
47class Net(Cell):
48    def __init__(self, w1, w2, strategy1=None, strategy2=None):
49        super().__init__()
50        self.less = P.Less().shard(strategy1)
51        self.w1 = Parameter(w1, "w1")
52        self.w2 = Parameter(w2, "w2")
53        self.select = P.Select().shard(strategy2)
54
55    def construct(self, x, b):
56        out = self.less(x, b)
57        out = self.select(out, self.w1, self.w2)
58        return out
59
60
61_x = Tensor(np.ones([16, 64, 32]), dtype=ms.float32)
62_b = Tensor(np.ones([16, 64, 32]), dtype=ms.float32)
63_w1 = Tensor(np.ones([128, 64, 32]), dtype=ms.float32)
64_w2 = Tensor(np.ones([128, 64, 32]), dtype=ms.float32)
65
66
67def compile_net(net):
68    learning_rate = 0.1
69    momentum = 0.9
70    epoch_size = 2
71    dataset = Dataset(_x, _b)
72    opt = Momentum(net.trainable_params(), learning_rate, momentum)
73    model = Model(net, optimizer=opt)
74    model.train(epoch_size, dataset, dataset_sink_mode=False)
75    context.reset_auto_parallel_context()
76
77
78def test_select_data_parallel():
79    context.set_auto_parallel_context(
80        parallel_mode="semi_auto_parallel", device_num=8, global_rank=0)
81    strategy1 = ((8, 1, 1), (8, 1, 1))
82    strategy2 = ((8, 1, 1), (8, 1, 1), (8, 1, 1))
83    net = Net(_w1, _w2, strategy1, strategy2)
84    compile_net(net)
85
86
87def test_select_model_parallel():
88    context.set_auto_parallel_context(
89        parallel_mode="semi_auto_parallel", device_num=8, global_rank=0)
90    strategy1 = ((2, 2, 2), (2, 2, 2))
91    strategy2 = ((2, 2, 2), (2, 2, 2), (2, 2, 2))
92    net = Net(_w1, _w2, strategy1, strategy2)
93    compile_net(net)
94
95
96def test_select_mirror():
97    context.set_auto_parallel_context(
98        parallel_mode="semi_auto_parallel", device_num=8, global_rank=0)
99    strategy1 = ((1, 2, 2), (1, 2, 2))
100    strategy2 = ((1, 2, 2), (1, 2, 2), (1, 2, 2))
101    net = Net(_w1, _w2, strategy1, strategy2)
102    compile_net(net)
103
104
105def test_select_auto_parallel():
106    context.set_auto_parallel_context(
107        parallel_mode="auto_parallel", device_num=8, global_rank=0)
108    net = Net(_w1, _w2)
109    compile_net(net)
110
111
112def test_select_strategy_error():
113    context.set_auto_parallel_context(
114        parallel_mode="semi_auto_parallel", device_num=8, global_rank=0)
115    strategy1 = ((2, 2, 2), (2, 2, 2))
116    strategy2 = ((8, 1, 1), (2, 2, 2), (2, 2, 2))
117    net = Net(_w1, _w2, strategy1, strategy2)
118    with pytest.raises(RuntimeError):
119        compile_net(net)
120