Deep-learning-based-surrogate-model/args.py at main · GetRed/Deep-learning-based-surrogate-model · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
#%%
import os
import argparse
import torch
#%%
class UNetParser(argparse.ArgumentParser):
    def __init__(self):
        super(UNetParser, self).__init__(description='Dense Encoder-Decoder Convolutional Network')

        self.add_argument('--blocks', type=list, default=(5, 10, 5), help='list of number of layers in each block in decoding net')
        self.add_argument('--growth-rate', type=int, default=40, help='output of each conv')
        self.add_argument('--drop-rate', type=float, default=0, help='dropout rate')
        self.add_argument('--bn-size', type=int, default=8, help='bottleneck size: bn_size * growth_rate')
        self.add_argument('--bottleneck', action='store_true', default=False, help='enable bottleneck in the dense blocks')
        self.add_argument('--init-features', type=int, default=48, help='# initial features after the first conv layer')

        #self.add_argument('--rundir', type=str, default='/home/sunrc/work/VICcases/huaihe_new/calibdL/', help='root dir to train')
        self.add_argument('--rundir', type=str, default='/home/shengsy/sun2023/', help='root dir to train')
        self.add_argument('--epochs', type=int, default=200, help='number of epochs to train')
        self.add_argument('--lr', type=float, default=0.0001, help='learnign rate')
        self.add_argument('--batch-size', type=int, default=400, help='batch size')
        self.add_argument('--batch-size-valid', type=int, default=400, help='validation batch size')
        self.add_argument('--patience', type=int, default=7, help='how long to wait after last time validation loss improved')
        self.add_argument('--epoch-save', type=int, default=5, help='number of epochs to save model')

    def parse(self):
        # cfg = self.parse_args()
        cfg = self.parse_known_args()[0]

        cfg.device = 'cuda:0' if torch.cuda.is_available() else 'cpu'
        cfg.nbatch_valid = 400

        run_name = f'surro_run_batch{cfg.batch_size}/'
        cfg.input_path = cfg.rundir + 'input_new/'
        # cfg.input_path = cfg.rundir + 'input/'
        # cfg.out_dir = cfg.rundir + run_name
        cfg.out_dir = '/home/shengsy/sun2023/surro_dL/HPCresults/UNet/AE_7pnewin/'
        # cfg.out_dir = '/home/sunrc/work/VICcases/huaihe_new/surrogatedL/HPCresults/paper_ARnet/sample_400/'
        cfg.out_model_dir = cfg.rundir + run_name + 'model/'
        cfg.tensorb_dir = cfg.rundir + run_name + 'tb/'

        if os.path.exists(cfg.out_dir) is False:
            os.mkdir(cfg.out_dir)
            os.mkdir(cfg.out_model_dir)
            os.mkdir(cfg.tensorb_dir)

        return cfg
# %%
class ResNetParser(argparse.ArgumentParser):
    def __init__(self):
        super(ResNetParser, self).__init__(description='Dense Encoder-Decoder Convolutional Network')

        self.add_argument('--blocks', type=list, default=(5, 10, 5), help='list of number of layers in each block in decoding net')
        self.add_argument('--growth-rate', type=int, default=40, help='output of each conv')
        self.add_argument('--drop-rate', type=float, default=0, help='dropout rate')
        self.add_argument('--bn-size', type=int, default=8, help='bottleneck size: bn_size * growth_rate')
        self.add_argument('--bottleneck', action='store_true', default=False, help='enable bottleneck in the dense blocks')
        self.add_argument('--init-features', type=int, default=48, help='# initial features after the first conv layer')

        #self.add_argument('--rundir', type=str, default='/home/sunrc/work/VICcases/huaihe_new/calibdL/', help='root dir to train')
        self.add_argument('--rundir', type=str, default='/home/shengsy/sun2023/', help='root dir to train')
        self.add_argument('--epochs', type=int, default=200, help='number of epochs to train')
        self.add_argument('--lr', type=float, default=0.0001, help='learnign rate')
        self.add_argument('--batch-size', type=int, default=400, help='batch size')
        self.add_argument('--batch-size-valid', type=int, default=400, help='validation batch size')
        self.add_argument('--patience', type=int, default=7, help='how long to wait after last time validation loss improved')
        self.add_argument('--epoch-save', type=int, default=5, help='number of epochs to save model')

    def parse(self):
        # cfg = self.parse_args()
        cfg = self.parse_known_args()[0]

        cfg.device = 'cuda:0' if torch.cuda.is_available() else 'cpu'
        cfg.nbatch_valid = 400

        run_name = f'resnet_run_batch{cfg.batch_size}/'
        cfg.input_path = cfg.rundir + 'input_new/'
        # cfg.input_path = cfg.rundir + 'input/'
        # cfg.out_dir = cfg.rundir + run_name
        cfg.out_dir = '/home/shengsy/sun2023/surro_dL/HPCresults/ResNet/AE_7pnewin/'
        # cfg.out_dir = '/home/sunrc/work/VICcases/huaihe_new/surrogatedL/HPCresults/paper_ARnet/sample_400/'
        cfg.out_model_dir = cfg.rundir + run_name + 'model/'
        cfg.tensorb_dir = cfg.rundir + run_name + 'tb/'

        if os.path.exists(cfg.out_dir) is False:
            os.mkdir(cfg.out_dir)
            os.mkdir(cfg.out_model_dir)
            os.mkdir(cfg.tensorb_dir)

        return cfg
# %%
class ARnetParser(argparse.ArgumentParser):
    def __init__(self):
        super(ARnetParser, self).__init__(description='Dense Encoder-Decoder Convolutional Network')

        self.add_argument('--blocks', type=list, default=(5, 10, 5), help='list of number of layers in each block in decoding net')
        self.add_argument('--growth-rate', type=int, default=40, help='output of each conv')
        self.add_argument('--drop-rate', type=float, default=0, help='dropout rate')
        self.add_argument('--bn-size', type=int, default=8, help='bottleneck size: bn_size * growth_rate')
        self.add_argument('--bottleneck', action='store_true', default=False, help='enable bottleneck in the dense blocks')
        self.add_argument('--init-features', type=int, default=48, help='# initial features after the first conv layer')

        # self.add_argument('--rundir', type=str, default='/home/sunrc/work/VICcases/huaihe_new/calibdL/', help='root dir to train')
        self.add_argument('--rundir', type=str, default='/home/shengsy/sun2023/', help='root dir to train')
        self.add_argument('--epochs', type=int, default=200, help='number of epochs to train')
        self.add_argument('--lr', type=float, default=0.0001, help='learnign rate')
        self.add_argument('--batch-size', type=int, default=400, help='batch size')
        self.add_argument('--batch-size-valid', type=int, default=400, help='validation batch size')
        self.add_argument('--patience', type=int, default=7, help='how long to wait after last time validation loss improved')
        self.add_argument('--epoch-save', type=int, default=5, help='number of epochs to save model')

    def parse(self):
        # cfg = self.parse_args()
        cfg = self.parse_known_args()[0]

        cfg.device = 'cuda:0' if torch.cuda.is_available() else 'cpu'
        cfg.nbatch_valid = 400

        run_name = f'arnet_run_batch{cfg.batch_size}/'
        cfg.input_path = cfg.rundir + 'input_new/'
        # cfg.input_path = cfg.rundir + 'input/'
        # cfg.out_dir = cfg.rundir + run_name
        cfg.out_dir = '/home/shengsy/sun2023/surro_dL/HPCresults/ARnet/AE_7pnewin/'
        # cfg.out_dir = '/home/sunrc/work/VICcases/huaihe_new/surrogatedL/HPCresults/paper_ARnet/sample_400/'
        cfg.out_model_dir = cfg.rundir + run_name + 'model/'
        cfg.tensorb_dir = cfg.rundir + run_name + 'tb/'

        if os.path.exists(cfg.out_dir) is False:
            os.mkdir(cfg.out_dir)
            os.mkdir(cfg.out_model_dir)
            os.mkdir(cfg.tensorb_dir)

        return cfg
#%%
#class FSTRParser(argparse.ArgumentParser):
#    def __init__(self):
#        super(FSTRParser, self).__init__(description='FSTR (Forced Spatio-Temporal RNN) Network')
#
#        # FSTR模型参数
#        self.add_argument('--num-layers', type=int, default=3, help='number of LSTM layers')
#        self.add_argument('--num-hidden', type=list, default=[64, 64, 64], help='number of hidden channels in each layer')
#        self.add_argument('--filter-size', type=int, default=5, help='filter size for convolutions')
#        self.add_argument('--stride', type=int, default=1, help='stride for convolutions')
#
#        # 通道数参数（会根据数据集自动设置）
#        self.add_argument('--act-channel', type=int, default=7, help='channel for forcings')
#        self.add_argument('--static-channel', type=int, default=7, help='channel for static parameters')
#        self.add_argument('--init-cond-channel', type=int, default=1, help='channel for initial condition')
#        self.add_argument('--out-channel', type=int, default=1, help='output channel')
#
#        # 训练参数
#        self.add_argument('--rundir', type=str, default='/home/shengsy/sun2023/', help='root dir to train')
#        self.add_argument('--epochs', type=int, default=200, help='number of epochs to train')
#        self.add_argument('--lr', type=float, default=0.0001, help='learning rate')
#        self.add_argument('--batch-size', type=int, default=400, help='batch size')
#        self.add_argument('--batch-size-valid', type=int, default=400, help='validation batch size')
#        self.add_argument('--patience', type=int, default=7, help='how long to wait after last time validation loss improved')
#        self.add_argument('--epoch-save', type=int, default=5, help='number of epochs to save model')
#
#    def parse(self):
#        cfg = self.parse_known_args()[0]
#
#        cfg.device = 'cuda:0' if torch.cuda.is_available() else 'cpu'
#        cfg.nbatch_valid = 400
#
#        run_name = f'fstr_run_batch{cfg.batch_size}/'
#        cfg.input_path = cfg.rundir + 'input_new/'
#        cfg.out_dir = '/home/shengsy/sun2023/surro_dL/HPCresults/FSTR/AE_7pnewin/'
#        cfg.out_model_dir = cfg.rundir + run_name + 'model/'
#        cfg.tensorb_dir = cfg.rundir + run_name + 'tb/'
#
#        if os.path.exists(cfg.out_dir) is False:
#            os.makedirs(cfg.out_dir, exist_ok=True)
#            os.makedirs(cfg.out_model_dir, exist_ok=True)
#            os.makedirs(cfg.tensorb_dir, exist_ok=True)
#
#        return cfg
#%%
class FSTRParser(argparse.ArgumentParser):
    def __init__(self):
        super(FSTRParser, self).__init__(description='FSTR Network', allow_abbrev=False)

        self.add_argument('--rundir', type=str, default='/home/shengsy/sun2023/', help='root dir to train')
        self.add_argument('--epochs', type=int, default=200, help='number of epochs to train')
        self.add_argument('--lr', type=float, default=0.0001, help='learnign rate')
        self.add_argument('--batch-size', type=int, default=400, help='batch size')
        self.add_argument('--batch-size-valid', type=int, default=400, help='validation batch size')
        self.add_argument('--frames_input', type=int, default=10, help='sum of input frames')
        self.add_argument('--frames_output', type=int, default=10, help='sum of predict frames')
        self.add_argument('--patience', type=int, default=7, help='how long to wait after last time validation loss improved')
        self.add_argument('--epoch-save', type=int, default=5, help='number of epochs to save model')

    def parse(self):
        # cfg = self.parse_args()
        cfg = self.parse_known_args()[0]

        cfg.device = 'cuda:0' if torch.cuda.is_available() else 'cpu'
        cfg.nbatch_valid = 400

        run_name = f'fstr_run_batch{cfg.batch_size}/'
        cfg.input_path = cfg.rundir + 'input_new/'
        cfg.out_dir = '/home/shengsy/sun2023/surro_dL/HPCresults/FSTR/AE_7pnewin/'
        cfg.out_model_dir = cfg.rundir + run_name + 'model/'
        cfg.tensorb_dir = cfg.rundir + run_name + 'tb/'

        if os.path.exists(cfg.out_dir) is False:
            os.makedirs(cfg.out_dir, exist_ok=True)
            os.makedirs(cfg.out_model_dir, exist_ok=True)
            os.makedirs(cfg.tensorb_dir, exist_ok=True)

        return cfg
#%%