-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy patheval_fg.py
154 lines (132 loc) · 6.83 KB
/
eval_fg.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
import argparse
import os
import time
import torch
import torch.optim as optim
import torch.nn as nn
import tqdm
from torch.utils.data import DataLoader
from data_load.DataSets.MiniImageNet_fg import *
from method.good_metric import Net
from method.local_match import *
from method.stl_deepbdc import *
from data_load.transform_cfg import *
import pprint
DATA_DIR = ''
_utils_pp = pprint.PrettyPrinter()
def pprint(x):
_utils_pp.pprint(x)
def parse_option():
parser = argparse.ArgumentParser('arguments for model pre-train')
# about dataset and network
parser.add_argument('--dataset', type=str, default='miniimagenet',
choices=['miniimagenet', 'cub', 'tieredimagenet', 'fc100', 'tieredimagenet_yao', 'cifar_fs'])
parser.add_argument('--data_root', type=str, default=DATA_DIR)
parser.add_argument('--transform', type=str, default='A', choices=transforms_list)
# about model :
parser.add_argument('--drop_gama', default=0.5, type= float)
parser.add_argument('--MLP_2', default=False, action='store_true')
parser.add_argument("--beta", default=0.01, type=float)
parser.add_argument('--drop_rate', default=0.5, type=float)
parser.add_argument('--reduce_dim', default=128, type=int)
parser.add_argument('--idea', default='a+-b', choices=['ab', 'a+-b', 'bdc'])
parser.add_argument('--FPN_list', default=None, nargs='+', type=int)
parser.add_argument('--flatten_fpn', default=False, action='store_true')
# about meta test
parser.add_argument('--val_freq',default=5,type=int)
parser.add_argument('--local_mode',default='local_mix', choices=['cell', 'local_mix' ,'cell_mix','mask_pool'])
parser.add_argument('--set', type=str, default='val', choices=['val', 'test'], help='the set for validation')
parser.add_argument('--n_way', type=int, default=5)
parser.add_argument('--n_shot', type=int, default=1)
parser.add_argument('--n_aug_support_samples',type=int, default=1)
parser.add_argument('--n_queries', type=int, default=15)
parser.add_argument('--temperature', type=float, default=12.5)
parser.add_argument('--metric', type=str, default='cosine')
parser.add_argument('--n_episodes', type=int, default=1000)
parser.add_argument('--n_local_proto', default=3, type=int)
parser.add_argument('--num_workers', default=2, type=int)
# test_batch_size is 1 maen 1 episode of fsl
parser.add_argument('--test_batch_size',default=1)
parser.add_argument('--sfc_lr', default=100,type = float)
parser.add_argument('--sfc_bs', default=5, type=int)
parser.add_argument('--sfc_update_step', default=100)
parser.add_argument('--include_bg', default=False, action='store_true')
# parser.add_argument('--norm',default='center')
# about deepemd setting
parser.add_argument('--norm', type=str, default='center', choices=['center'])
parser.add_argument('--solver', type=str, default='opencv', choices=['opencv'])
parser.add_argument('--deep_emd', default=False, action='store_true')
parser.add_argument('-form', type=str, default='L2', choices=['QP', 'L2'])
parser.add_argument('-l2_strength', type=float, default=0.000001)
# setting
parser.add_argument('--gpu', default=0, type=int)
parser.add_argument('--save_dir', default='checkpoint')
parser.add_argument('--continue_pretrain',default=False,action='store_true')
parser.add_argument('--test_LR', default=False, action='store_true')
parser.add_argument('--model_id', default=None, type=str)
parser.add_argument('--model_type',default='best',choices=['best','last'])
parser.add_argument('--seed', default=1, type=int)
parser.add_argument('--no_save_model', default=False, action='store_true')
parser.add_argument('--feature_pyramid', default=False, action='store_true')
parser.add_argument('--method', default='local_proto', choices=['local_proto', 'good_metric', 'stl_deepbdc'])
parser.add_argument('--distill_model', default=None,type=str,help='about distillation model path')
parser.add_argument('--penalty_c', default=1.0, type=float)
parser.add_argument('--stop_grad', default=False, action='store_true')
parser.add_argument('--learnable_alpha', default=False, action='store_true')
parser.add_argument('--idea_variant', default=False, action='store_true')
parser.add_argument('--normalize_feat', default=False, action='store_true')
parser.add_argument('--fg_extract', default=False, action='store_true')
args = parser.parse_args()
if args.deep_emd:
args.method = 'deep_emd'
return args
def model_load(args,model):
# method = 'deep_emd' if args.deep_emd else 'local_match'
method = args.method
save_path = os.path.join(args.save_dir, args.dataset + "_" + method + "_resnet12_"+args.model_type
+ ("_"+str(args.model_id) if args.model_id else "") + ".pth")
if args.distill_model is not None:
save_path = os.path.join(args.save_dir, args.distill_model)
print('teacher model path: ' + save_path)
state_dict = torch.load(save_path)['model']
model.load_state_dict(state_dict)
return model
def main():
args = parse_option()
pprint(args)
if args.gpu:
gpu_device = str(args.gpu)
else:
gpu_device = "0"
os.environ['CUDA_VISIBLE_DEVICES'] = gpu_device
if args.dataset == 'miniimagenet':
train_trans, test_trans = transforms_options[args.transform]
val_sup_trans = test_trans if args.n_aug_support_samples == 1 else train_trans
meta_test_loader = DataLoader(MetaImageNet(args=args, partition='test',
train_transform=val_sup_trans,
test_transform=test_trans),
batch_size=args.test_batch_size, shuffle=False, drop_last=False,
num_workers=args.num_workers)
num_cls = 64
if args.method in [ 'local_proto','deep_emd']:
model = Local_match(args,num_classes=num_cls,local_proto=args.n_local_proto,mask_ad=True).cuda()
elif args.method in ['stl_deepbdc']:
model = stl_deepbdc(args, num_classes=num_cls).cuda()
elif args.method in ['good_metric']:
model = Net(args, num_classes=num_cls, ).cuda()
else:
model = None
assert model != None
if args.continue_pretrain:
model = model_load(args,model)
print("-"*20+" start meta test... "+"-"*20)
model.eval()
# gen_test = tqdm.tqdm(meta_test_loader)
with torch.no_grad():
tic = time.time()
mean, confidence = model.meta_test_loop(meta_test_loader)
print()
print("meta_val acc: {:.2f} +- {:.2f} elapse: {:.2f} min".format(mean * 100, confidence * 100,
(time.time() - tic) / 60))
if __name__ == '__main__':
main()