master-thesis/arguments.py at main · maruoss/master-thesis · 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
from datamodule import Dataset, DataModule
from model.neuralnetwork import FFN
from model.transformer import TransformerEncoder

import pdb


def load_args(parser, mode: str, model: str):
    """load arguments for respective mode and model"""
    globals()[f"args_{model}"](parser)


def args_nn(parser):
    # EarlyStopping
    group = parser.add_argument_group("Early Stopping Configuration")
    # group.add_argument("--monitor", type=str, default="loss/val_loss")
    # group.add_argument("--es_mode", type=str, default="min")
    group.add_argument("--patience", type=int, default=5, #multiply with check_val_every
                        help="number of bad validation epochs before stop, depends "
                        "on 'check_val_every' parameter as well")

    # ModelCheckpoint
    # group = parser.add_argument_group("Model Checkpoint Configuration")
    # group.add_argument("--monitor", type=str, default="loss/val_loss")
    # group.add_argument("--save_top_k", type=int, default=1)
    # group.add_argument("--check_mode", type=str, default="min")

    # dm
    # already implemented in model method: group = parser.add_argument_group("Data Module Configuration")
    group = DataModule.add_model_specific_args(parser)  #add additional arguments directly in class method

    # model
    # already implemented in model method: group = parser.add_argument_group("Model Configuration")
    group = FFN.add_model_specific_args(parser) #add additional arguments directly in class

    # trainer
    group = parser.add_argument_group("Training Configuration")
    group.add_argument("--max_epochs", type=int, default=1000)
    group.add_argument("--check_val_every", type=int, default=10)
    # group.add_argument("--cpus_per_trial", type=int, default=8)


def args_lin(parser):
    # Dataset args
    group = Dataset.add_model_specific_args(parser)

    # group = parser_train.add_argument_group("Model Configuration")
    group.add_argument("--loss", type=str, choices=["log_loss", "hinge"],
                        required=True)
    group.add_argument("--pca", action="store_true")

    group = parser.add_argument_group("Sklearn Tune Configuration")
    group.add_argument("--max_iters", type=int, default=300)
    # group.add_argument("--njobs", type=int, default=2) #how many trials in parallel


def args_svm(parser):
    # Dataset args
    group = Dataset.add_model_specific_args(parser)

    # group = parser_train.add_argument_group("Model Configuration")
    # group.add_argument("--loss", type=str, choices=["hinge", "log_loss"],
    #                     default="log_loss")
    group.add_argument("--pca", action="store_true")

    group = parser.add_argument_group("Sklearn Tune Configuration")
    group.add_argument("--max_iters", type=int, default=300)
    # group.add_argument("--njobs", type=int, default=2) #how many trials in parallel


def args_rf(parser):
    # Dataset args
    group = Dataset.add_model_specific_args(parser)

    # group = parser_train.add_argument_group("Model Configuration")
    # group.add_argument("--loss", type=str, choices=["hinge", "log_loss"],
    #                     default="log_loss")
    group.add_argument("--pca", action="store_true")

    group = parser.add_argument_group("Sklearn Tune Configuration")
    group.add_argument("--n_estimators", type=int, default=100)
    # group.add_argument("--njobs", type=int, default=2) #how many trials in parallel


def args_xgb(parser):
    # Dataset args
    group = Dataset.add_model_specific_args(parser)

    # EarlyStopping
    group = parser.add_argument_group("Early Stopping Configuration")
    # group.add_argument("--monitor", type=str, default="loss/val_loss")
    # group.add_argument("--es_mode", type=str, default="min")
    group.add_argument("--patience", type=int, default=32,
                            help="number of bad epochs before stop")

    group = parser.add_argument_group("Training Configuration")
    group.add_argument("--num_boost_round", type=int, default=1000)
    # group.add_argument("--cpus_per_trial", type=int, default=8)


def args_transformer(parser):
    # EarlyStopping
    group = parser.add_argument_group("Early Stopping Configuration")
    # group.add_argument("--monitor", type=str, default="loss/val_loss")
    # group.add_argument("--es_mode", type=str, default="min")
    group.add_argument("--patience", type=int, default=5, #multiply with check_val_every
                        help="number of bad validation epochs before stop, depends "
                        "on 'check_val_every' parameter as well")

    # ModelCheckpoint
    # group = parser.add_argument_group("Model Checkpoint Configuration")
    # group.add_argument("--monitor", type=str, default="loss/val_loss")
    # group.add_argument("--save_top_k", type=int, default=1)
    # group.add_argument("--check_mode", type=str, default="min")

    # dm
    # already implemented in model method: group = parser.add_argument_group("Data Module Configuration")
    group = DataModule.add_model_specific_args(parser)  #add additional arguments directly in class method

    # model
    # already implemented in model method: group = parser.add_argument_group("Model Configuration")
    group = TransformerEncoder.add_model_specific_args(parser) #add additional arguments directly in class

    # trainer
    group = parser.add_argument_group("Training Configuration")
    group.add_argument("--max_epochs", type=int, default=1000)
    group.add_argument("--check_val_every", type=int, default=10)
    # group.add_argument("--cpus_per_trial", type=int, default=8)