-
Notifications
You must be signed in to change notification settings - Fork 0
/
run.py
68 lines (57 loc) · 1.46 KB
/
run.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
"""run file
do not track with git
"""
import os
import torch
from bird_cloud_gnn.cross_validation import (
get_dataloaders,
kfold_evaluate,
leave_one_origin_out_evaluate,
)
from bird_cloud_gnn.gnn_model import GCN
from bird_cloud_gnn.radar_dataset import RadarDataset
from dgl.dataloading import GraphDataLoader
from torch.utils.data.sampler import SubsetRandomSampler
import torch.nn.functional as F
from torch import nn
DATA_PATH = "../data/volume_2/parquet/"
os.listdir(DATA_PATH)
features = [
"range",
"azimuth",
"elevation",
"x",
"y",
"z",
"DBZH",
"DBZV",
"TH",
"TV",
"PHIDP",
"RHOHV",
]
dataset = RadarDataset(
data=DATA_PATH,
features=features,
target="BIOLOGY",
num_nodes=20,
max_poi_per_label=20,
max_edge_distance=5_000.0,
)
num_examples = len(dataset)
num_train = int(num_examples * 0.8)
train_idx = torch.arange(num_train)
test_idx = torch.arange(num_train, num_examples)
DATA_PATH = "../data/volume_2/parquet/"
model = GCN(len(dataset.features), [(16, nn.ReLU()), (16, nn.ReLU()), (2, None)])
train_dataloader, test_dataloader = get_dataloaders(
dataset, train_idx, test_idx, batch_size=512
)
model.oneline_description()
model.fit(train_dataloader)
model.evaluate(test_dataloader)
print("Model's state_dict:")
for param_tensor in model.state_dict():
print(param_tensor, "\t", model.state_dict()[param_tensor])
# kfold_evaluate(dataset)
# leave_one_origin_out_evaluate(dataset)