-
Notifications
You must be signed in to change notification settings - Fork 0
/
hubconf.py
160 lines (120 loc) · 5.26 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
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
#encoding=utf8
# Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
'''
hubconf.py for pytorch_gan_zoo repo
## Users can get the diverse models of pytorch_gan_zoo by calling
hub_model = hub.load(
'facebookresearch/pytorch_gan_zoo:master',
$MODEL_NAME, #
config = None,
useGPU = True,
pretrained=False) # (Not pretrained models online yet)
Available model'names are [DCGAN, PGAN, StyleGAN].
The config option should be a dictionnary defining the training parameters of
the model. See ??/pytorch_gan_zoo/models/trainer/standard_configurations to see
all possible options
## How can I use my model ?
### Build a random vector
inputRandom, randomLabels = model.buildNoiseData((int) $BATCH_SIZE)
### Feed a random vector to the model
model.test(inputRandom,
getAvG=True,
toCPU=True)
Arguments:
- getAvG (bool) get the smoothed version of the generator (advised)
- toCPU (bool) if set to False the output tensor will be a torch.cuda tensor
### Acces the generator
model.netG()
### Acces the discriminator
model.netD()
## Can I train my model ?
Of course. You can set all training parameters in the constructor (losses to use,
learning rate, number of iterations etc...) and use the optimizeParameters()
method to make a training steps.
Typically here will be a sample code:
for input_real in dataset:
allLosses = model.optimizeParameters(inputs_real)
# Do something with the losses
Please have a look at
models/trainer/standard_configurations to see all the
training parameters you can use.
'''
import torch
import torch.utils.model_zoo as model_zoo
# Optional list of dependencies required by the package
dependencies = ['torch']
def PGAN(pretrained=False, *args, **kwargs):
"""
Progressive growing model
pretrained (bool): load a pretrained model ?
model_name (string): if pretrained, load one of the following models
celebaHQ-256, celebaHQ-512, DTD, celeba, cifar10. Default is celebaHQ.
"""
from models.progressive_gan import ProgressiveGAN as PGAN
if 'config' not in kwargs or kwargs['config'] is None:
kwargs['config'] = {}
model = PGAN(useGPU=kwargs.get('useGPU', True),
storeAVG=True,
**kwargs['config'])
checkpoint = {"celebAHQ-256": 'https://dl.fbaipublicfiles.com/gan_zoo/PGAN/celebaHQ_s6_i80000-6196db68.pth',
"celebAHQ-512": 'https://dl.fbaipublicfiles.com/gan_zoo/PGAN/celebaHQ16_december_s7_i96000-9c72988c.pth',
"DTD": 'https://dl.fbaipublicfiles.com/gan_zoo/PGAN/testDTD_s5_i96000-04efa39f.pth',
"celeba": "https://dl.fbaipublicfiles.com/gan_zoo/PGAN/celebaCropped_s5_i83000-2b0acc76.pth"}
if pretrained:
if "model_name" in kwargs:
if kwargs["model_name"] not in checkpoint.keys():
raise ValueError("model_name should be in "
+ str(checkpoint.keys()))
else:
print("Loading default model : celeba")
kwargs["model_name"] = "celeba"
# # 加载预训练模型:
# state_dict = model_zoo.load_url(checkpoint[kwargs["model_name"]],
# map_location='cpu')
state_dict = torch.load(r'./test_Nets/celebaCropped_s5_i83000-2b0acc76.pth')
model.load_state_dict(state_dict)
return model
def StyleGAN(pretrained=False, *args, **kwargs):
"""
NVIDIA StyleGAN
pretrained (bool): load a 1024x1024 model trained on FlickrHQ
"""
from models.styleGAN import StyleGAN
if 'config' not in kwargs or kwargs['config'] is None:
kwargs['config'] = {}
model = StyleGAN(useGPU=kwargs.get('useGPU', True),
storeAVG=True,
**kwargs['config'])
checkpoint = 'https://dl.fbaipublicfiles.com/gan_zoo/StyleGAN/FFHQ_styleGAN-7cbdec00.pth'
if pretrained:
print("Loading default model : Flickr-HQ")
state_dict = model_zoo.load_url(checkpoint,
map_location='cpu')
model.load_state_dict(state_dict)
return model
def DCGAN(pretrained=False, *args, **kwargs):
"""
DCGAN basic model
pretrained (bool): load a pretrained model ? In this case load a model
trained on fashionGen cloth
"""
from models.DCGAN import DCGAN
if 'config' not in kwargs or kwargs['config'] is None:
kwargs['config'] = {}
model = DCGAN(useGPU=kwargs.get('useGPU', True),
storeAVG=True,
**kwargs['config'])
checkpoint = 'https://dl.fbaipublicfiles.com/gan_zoo/DCGAN_fashionGen-1d67302.pth'
if pretrained:
state_dict = model_zoo.load_url(checkpoint, map_location='cpu')
model.load_state_dict(state_dict)
return model