-
Notifications
You must be signed in to change notification settings - Fork 0
/
predict.py
56 lines (45 loc) · 1.25 KB
/
predict.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
import sys
import importlib
from train import ModelLightning
import models
import torch
import datasets
import pytorch_lightning as pyl
import numpy as np
import os
import sklearn.metrics
if __name__ == '__main__':
config_file = sys.argv[1]
ckpt_file = sys.argv[2]
output_file = sys.argv[3]
config = importlib.import_module(config_file).config
backbone = models.HierarchicalTransformer(config)
model = ModelLightning(
config, backbone=backbone)
model.load_state_dict(torch.load(ckpt_file)['state_dict'])
model.eval()
dataset = datasets.DygDatasetTest(config, 'test')
dataloader = torch.utils.data.DataLoader(
dataset=dataset,
batch_size=config['batch_size'],
shuffle=False,
num_workers=0,
collate_fn=datasets.dyg_test_collate_fn
)
trainer = pyl.Trainer(
gpus=1
)
with torch.no_grad():
pred = trainer.predict(
model, dataloader)
pass
pred = np.hstack(pred)
test_index = np.load(
os.path.join(config['dataset_path'], 'test_index.npy'))
# pred[test_index[:, 1]==-1] = 0
with open(output_file, 'w') as fout:
for p in pred:
fout.write(f'{p}\n')
pass
pass
pass