THIS-ODE/run_NODE.py at main · shib0li/THIS-ODE · 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
import numpy as np
import torch
from torchdiffeq import odeint
import torch.autograd.functional as F
import random

import torch.nn as nn
import os
import pickle
import time
import fire

from tqdm.auto import tqdm, trange
from torch.utils.data import Dataset, DataLoader

from core.model import *
from core.NODE import *
from core.NODE_noise import *
from core.NODE_auto import *

from data.real_events import EventData

from infrastructure.misc import *
from infrastructure.configs import *

np.random.seed(0)
torch.manual_seed(0)
random.seed(0)


def evaluation(**kwargs):

    config = NodeExpConfig()
    config.parse(kwargs)

    device = torch.device(config.device)

    domain = config.domain
    fold = config.fold

    dataset_train = EventData(domain, mode='train', fold=fold)
    dataset_test = EventData(domain, mode='test', fold=fold)

    if config.est == 'pt':
        method = 'NODE'
    elif config.est == 'auto':
        method = 'NODE_auto'
    else:
        method = 'NODE_noise'

    res_path = os.path.join(
        '__res__',
        dataset_train.domain,
        method,
        'rank'+str(config.R),
        'fold{}'.format(dataset_train.fold)
    )

    log_path = os.path.join(
        '__log__',
        dataset_train.domain,
        method,
        'rank'+str(config.R),
        'fold{}'.format(dataset_train.fold)
    )

    create_path(res_path)
    create_path(log_path)

    logger = get_logger(logpath=os.path.join(log_path, 'exp.log'), displaying=config.verbose)
    logger.info(config)

    batch_size = config.batch_size
    R = config.R

    if config.est == 'pt':
        node = NODE(
            nmod=dataset_train.nmod,
            nvec=dataset_train.nvec,
            R=R,
            batch_size=batch_size,
            int_steps=config.int_steps,
            nFF=config.nFF,
            solver=config.solver
        )
    elif config.est == 'auto':
        node = NODE_auto(
            nmod=dataset_train.nmod,
            nvec=dataset_train.nvec,
            R=R,
            nFF_init=config.nFF,
            nFF_dynamic=config.nFF,
            batch_size=batch_size,
            steps=config.int_steps,
            solver=config.solver
        )
    else:
        node = NODE_noise(
            nmod=dataset_train.nmod,
            nvec=dataset_train.nvec,
            R=R,
            batch_size=batch_size,
            int_steps=config.int_steps,
            nFF=config.nFF,
            solver=config.solver
        )

    node.todev(device)

    max_epochs = config.max_epochs
    learning_rate = config.learning_rate

    perform_meters = PerformMeters(save_path=res_path, logger=logger, test_interval=config.test_interval)

    node.train(dataset_train, dataset_test, max_epochs, learning_rate, perform_meters)


if __name__=='__main__':
    fire.Fire(evaluation)