• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1# Copyright 2020 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"""train_criteo."""
16import os
17import pytest
18
19from mindspore import context
20from mindspore.train.model import Model
21from mindspore.common import set_seed
22
23from src.deepfm import ModelBuilder, AUCMetric
24from src.config import DataConfig, ModelConfig, TrainConfig
25from src.dataset import create_dataset, DataType
26from src.callback import EvalCallBack, LossCallBack, TimeMonitor
27
28set_seed(1)
29
30@pytest.mark.level0
31@pytest.mark.platform_arm_ascend_training
32@pytest.mark.platform_x86_ascend_training
33@pytest.mark.env_onecard
34def test_deepfm():
35    data_config = DataConfig()
36    train_config = TrainConfig()
37    device_id = int(os.getenv('DEVICE_ID'))
38    context.set_context(mode=context.GRAPH_MODE, device_target="Ascend", device_id=device_id)
39    rank_size = None
40    rank_id = None
41
42    dataset_path = "/home/workspace/mindspore_dataset/criteo_data/mindrecord/"
43    print("dataset_path:", dataset_path)
44    ds_train = create_dataset(dataset_path,
45                              train_mode=True,
46                              epochs=1,
47                              batch_size=train_config.batch_size,
48                              data_type=DataType(data_config.data_format),
49                              rank_size=rank_size,
50                              rank_id=rank_id)
51
52    model_builder = ModelBuilder(ModelConfig, TrainConfig)
53    train_net, eval_net = model_builder.get_train_eval_net()
54    auc_metric = AUCMetric()
55    model = Model(train_net, eval_network=eval_net, metrics={"auc": auc_metric})
56
57    loss_file_name = './loss.log'
58    time_callback = TimeMonitor(data_size=ds_train.get_dataset_size())
59    loss_callback = LossCallBack(loss_file_path=loss_file_name)
60    callback_list = [time_callback, loss_callback]
61
62    eval_file_name = './auc.log'
63    ds_eval = create_dataset(dataset_path, train_mode=False,
64                             epochs=1,
65                             batch_size=train_config.batch_size,
66                             data_type=DataType(data_config.data_format))
67    eval_callback = EvalCallBack(model, ds_eval, auc_metric,
68                                 eval_file_path=eval_file_name)
69    callback_list.append(eval_callback)
70
71    print("train_config.train_epochs:", train_config.train_epochs)
72    model.train(train_config.train_epochs, ds_train, callbacks=callback_list)
73
74    export_loss_value = 0.52
75    print("loss_callback.loss:", loss_callback.loss)
76    assert loss_callback.loss < export_loss_value
77    export_per_step_time = 30.0
78    print("time_callback:", time_callback.per_step_time)
79    assert time_callback.per_step_time < export_per_step_time
80    print("*******test case pass!********")
81