forked from wimh966/QDrop
-
Notifications
You must be signed in to change notification settings - Fork 0
/
hubconf.py
85 lines (71 loc) · 2.87 KB
/
hubconf.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
from collections import OrderedDict
from QDrop.models.resnet import resnet18 as _resnet18
from QDrop.models.resnet import resnet50 as _resnet50
from QDrop.models.mobilenetv2 import mobilenetv2 as _mobilenetv2
from QDrop.models.mnasnet import mnasnet as _mnasnet
from QDrop.models.regnet import regnetx_600m as _regnetx_600m
from QDrop.models.regnet import regnetx_3200m as _regnetx_3200m
import torch
dependencies = ['torch']
prefix = '/mnt/lustre/weixiuying'
model_path = {
'resnet18': prefix+'/model_zoo/resnet18_imagenet.pth.tar',
'resnet50': prefix+'/model_zoo/resnet50_imagenet.pth.tar',
'mbv2': prefix+'/model_zoo/mobilenetv2.pth.tar',
'reg600m': prefix+'/model_zoo/regnet_600m.pth.tar',
'reg3200m': prefix+'/model_zoo/regnet_3200m.pth.tar',
'mnasnet': prefix+'/model_zoo/mnasnet.pth.tar',
'spring_resnet50': prefix+'/model_zoo/spring_resnet50.pth',
}
def resnet18(pretrained=False, **kwargs):
# Call the model, load pretrained weights
model = _resnet18(**kwargs)
if pretrained:
checkpoint = torch.load(model_path['resnet18'], map_location='cpu')
model.load_state_dict(checkpoint)
return model
def resnet50(pretrained=False, **kwargs):
# Call the model, load pretrained weights
model = _resnet50(**kwargs)
if pretrained:
checkpoint = torch.load(model_path['resnet50'], map_location='cpu')
model.load_state_dict(checkpoint)
return model
def spring_resnet50(pretrained=False, **kwargs):
# Call the model, load pretrained weights
model = _resnet50(**kwargs)
if pretrained:
checkpoint = torch.load(model_path['spring_resnet50'], map_location='cpu')
q = OrderedDict()
for k, v in checkpoint.items():
q[k[7:]] = v
model.load_state_dict(q)
return model
def mobilenetv2(pretrained=False, **kwargs):
# Call the model, load pretrained weights
model = _mobilenetv2(**kwargs)
if pretrained:
checkpoint = torch.load(model_path['mbv2'], map_location='cpu')
model.load_state_dict(checkpoint['model'])
return model
def regnetx_600m(pretrained=False, **kwargs):
# Call the model, load pretrained weights
model = _regnetx_600m(**kwargs)
if pretrained:
checkpoint = torch.load(model_path['reg600m'], map_location='cpu')
model.load_state_dict(checkpoint)
return model
def regnetx_3200m(pretrained=False, **kwargs):
# Call the model, load pretrained weights
model = _regnetx_3200m(**kwargs)
if pretrained:
checkpoint = torch.load(model_path['reg3200m'], map_location='cpu')
model.load_state_dict(checkpoint)
return model
def mnasnet(pretrained=False, **kwargs):
# Call the model, load pretrained weights
model = _mnasnet(**kwargs)
if pretrained:
checkpoint = torch.load(model_path['mnasnet'], map_location='cpu')
model.load_state_dict(checkpoint)
return model