/
parse_datasets.py
168 lines (147 loc) · 6.07 KB
/
parse_datasets.py
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
import os
import numpy as np
import torch
import torch.nn as nn
import importlib
import lib.utils as utils
from torch.utils.data import DataLoader
import sys
sys.path.append("../") # to access data_generators
from data_generators.me_ode_1d import MEODE1d
from data_generators.mujoco_physics import HopperPhysics
from data_generators.rotating_mnist import RotatingMnist
from data_generators.adni import Adni
from data_generators.tadpole import Tadpole
def parse_datasets(args, device):
def basic_collate_fn(batch,
time_steps,
args=args,
device=device,
data_type="train"):
if args.dataset == "adni":
batch = batch[0]
else:
batch = torch.stack(batch)
data_dict = {"data": batch, "time_steps": time_steps}
data_dict = utils.split_and_subsample_batch(data_dict,
args,
data_type=data_type)
return data_dict
dataset_name = args.dataset
data_gen = None
#n_total_tp = args.n_t + args.extrap * 2
#print("n_total_tp:", n_total_tp)
if dataset_name == "toy":
data_gen = MEODE1d(root="data",
download=False,
n_samples=args.n_samples,
n_t=args.n_t,
min_t=args.min_t,
max_t=args.max_t,
y0_mean=1.3,
y0_std=0.01,
fix_eff=0.3,
rand_eff_std=0.1,
device=device,
name="ME_ODE_1d")
elif dataset_name == "hopper":
# MuJoCo dataset
data_gen = HopperPhysics(root='data',
n_samples=args.n_samples,
n_t=args.n_t,
n_same_initial=1,
n_angles=args.n_angles,
fix_eff=0.3,
rand_eff_std=0.1,
device=device,
steps_to_skip=20,
name="ME_Hopper") # Generate data
elif dataset_name == "rotmnist":
# Rotating MNIST
data_gen = RotatingMnist(
root='data',
n_samples=args.n_samples,
n_t=args.n_t,
n_same_initial=4,
n_angles=args.n_angles,
frame_size=28,
device=device,
specific_digit=None if args.mnist_digit < 0 else args.mnist_digit,
n_styles=10,
mnist_data_path=None,
mnist_labels_path=None,
name="ME_Rotating_MNIST") # Generate data
elif dataset_name == "tadpole":
# Brain Images TADPOLE
data_gen = Tadpole(datapath='data/TADPOLE/cdrsb/',
device=device,
name="TADPOLE") # Generate data
elif dataset_name == "adni":
# Brain Images ADNI
adni_path = 'data/ADNI/MRI3_Seqs'
batch_size = args.batch_size
train_gen = Adni(datapath=adni_path,
device=device,
name="ADNI",
batch_size=batch_size,
is_train=True)
test_gen = Adni(datapath=adni_path,
device=device,
name="ADNI",
batch_size=batch_size,
is_train=False)
else:
raise Exception("Unknown dataset: %s" % dataset_name)
if dataset_name == "adni":
# if data_gen.n > args.n_samples:
# print("Loaded data set has %d samples" % data_gen.n)
# print("First %d samples will be chosen" % args.n_samples)
# dataset = dataset[:args.n_samples]
time_steps = train_gen.t
train_dataloader = DataLoader(
train_gen,
shuffle=False,
collate_fn=lambda batch: basic_collate_fn(
batch, time_steps, data_type="train"))
test_dataloader = DataLoader(test_gen,
shuffle=False,
collate_fn=lambda batch: basic_collate_fn(
batch, time_steps, data_type="test"))
data_objects = {
"data_gen": train_gen,
"train_dataloader": utils.inf_generator(train_dataloader),
"test_dataloader": utils.inf_generator(test_dataloader),
"input_dim": 1, # batch, time, dim/channel, ...
"n_train_batches": train_gen.n_train_batches,
"n_test_batches": train_gen.n_test_batches,
}
else:
dataset = data_gen.data
if len(dataset) > args.n_samples:
print("Loaded data set has %d samples" % len(dataset))
print("First %d samples will be chosen" % args.n_samples)
dataset = dataset[:args.n_samples]
time_steps = data_gen.t
train_y, test_y = utils.split_train_test(dataset, train_fraq=0.8)
batch_size = min(args.batch_size, args.n_samples)
train_dataloader = DataLoader(
train_y,
batch_size=batch_size,
shuffle=False,
collate_fn=lambda batch: basic_collate_fn(
batch, time_steps, data_type="train"))
test_dataloader = DataLoader(
test_y,
batch_size=batch_size, # args.n_samples,
shuffle=False,
collate_fn=lambda batch: basic_collate_fn(
batch, time_steps, data_type="test"))
data_objects = {
"data_gen": data_gen,
"train_dataloader": utils.inf_generator(train_dataloader),
"test_dataloader": utils.inf_generator(test_dataloader),
"input_dim": dataset.shape[2], # batch, time, dim/channel, ...
"n_train_batches": len(train_dataloader),
"n_test_batches": len(test_dataloader)
}
return data_objects