/
print_results.py
286 lines (249 loc) · 11.7 KB
/
print_results.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
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
import faiss # must be first imported in the main.py --> I don't know why
import os
import csv
import numpy as np
from tools.utils import get_save_dir, load_json
from configs import default
from models import backbone_names, head_names
import argparse
parser = argparse.ArgumentParser()
DATASETS = ['office31', 'officehome', 'visda', 'domainnet']
METHODS = ['SO', 'DANCE', 'OVANet', 'UniOT', 'WiSE-FT', 'ClipCrossModel', 'ClipZeroShot', 'ClipDistill', 'ClipDistillTemp1.0', 'AutoDistill']
# METHODS = ['SO', 'SO3.0', 'SO1.5', 'ClipDistill', 'ClipDistill3.0', 'ClipDistill1.5']
# METHODS += ['SOMarginP', 'ClipDistillMarginP']
# METHODS += ['FocalDistill']
METHODS += ['Auto_only_cal', 'Auto_wo_iid', 'Auto_wo_nll', 'Auto_wo_ood']
# METHODS = ['AutoDistill']
# METHODS += ['debug0.1', 'debug0.2', 'debug0.3', 'debug0.4', 'debug0.5', 'debug0.6', 'debug0.7', 'debug0.8', 'debug0.9', 'debug1.0'] # for hyperparameters analysis for CLIP distillation
DOMAINS = {'office31': ['amazon', 'dslr', 'webcam'],
'officehome': ['Art', 'Clipart', 'Product', 'RealWorld'],
'visda': ['syn', 'real'],
'domainnet': ['painting', 'real', 'sketch']}
NN = {'open-partial': {'office31': [10, 10],
'officehome': [10, 5],
'visda': [6, 3],
'domainnet': [150, 50]},
'open': {'office31': [10, 0],
'officehome': [15, 0],
'visda': [6, 0],
'domainnet': [150, 0]},
'closed': {'office31': [31, 0],
'officehome': [65, 0],
'visda': [12, 0],
'domainnet': [345, 0]},
'partial': {'office31': [10, 21],
'officehome': [25, 40],
'visda': [6, 6],
'domainnet': [150, 195]}
}
MAX_ITERS = {'open-partial': {'office31': 5000,
'officehome': 5000,
'visda': 10000,
'domainnet': 10000},
'open': {'office31': 5000,
'officehome': 5000,
'visda': 10000,
'domainnet': 10000},
'closed': {'office31': 10000,
'officehome': 10000,
'visda': 20000,
'domainnet': 20000},
'partial': {'office31': 10000,
'officehome': 10000,
'visda': 20000,
'domainnet': 20000},
}
# setting
STETTING = 'closed'
# ITERS = [1000 * (i+1) for i in range(10)]
ITERS = ['final']
SEEDS = [1, 2, 3]
RESULTS_TERMS = ['AA', 'H-score', 'H3-score', 'AUROC', 'OSCR']
if STETTING == 'closed' or STETTING == 'partial':
RESULTS_TERMS = ['OA', 'AA', 'Closed-set OA', 'Closed-set AA']
def main(args):
global METHODS, MAX_ITERS, STETTING
# args.backbone = 'dinov2_vitl14'
# args.backbone = 'resnet50'
# args.backbone = 'ViT-B/16'
# args.suffix = '_my'
save_result_dir = 'results'
if args.suffix is not None:
args.result_dir = args.result_dir + args.suffix
save_result_dir += args.suffix
result_dir_resnet50 = None
args.fixed_backbone = True
if not 'ViT' in args.backbone:
METHODS = ['SO', 'DANCE', 'OVANet', 'UniOT', 'ClipDistill', 'AutoDistill']
if args.backbone == 'resnet50':
args.fixed_backbone = False
assert STETTING == 'open-partial'
result_dir_resnet50 = '/data1/deng.bin/coding/uniood/experiments-0425'
overall_mean_steps = []
for step in ITERS:
overall_mean, all_columns_mean, all_columns_std = [], [], []
for dataset in DATASETS:
for method in METHODS:
# if method in ['SOMarginP', 'ClipDistillMarginP']:
# args.result_dir = './experimentsMarginP'
# else:
# args.result_dir = './experiments'
max_iter = MAX_ITERS[STETTING][dataset]
if args.backbone == 'resnet50' and method in ('SO', 'DANCE', 'OVANet', 'UniOT'):
max_iter = 10000
classifier_head = args.classifier_head
result_data_method = {}
for source_domain in DOMAINS[dataset]:
for target_domain in DOMAINS[dataset]:
if source_domain != target_domain and not (source_domain == 'real' and target_domain == 'syn'):
all_seed_dict = {}
for seed in SEEDS:
if method == 'ClipZeroShot':
seed_ = 1
else:
seed_ = seed
save_dir = get_save_dir(result_dir_resnet50 if (result_dir_resnet50 is not None and method in ('SO', 'DANCE', 'OVANet', 'UniOT')) else args.result_dir,
dataset,
method,
source_domain,
target_domain,
NN[STETTING][dataset][0],
NN[STETTING][dataset][1],
args.backbone,
args.optimizer,
args.base_lr,
classifier_head,
args.fixed_backbone,
args.fixed_BN,
args.image_augmentation,
args.batch_size,
f'{step}-{max_iter}',
seed_)
result = load_json(save_dir)
if result is not None:
all_seed_dict[seed] = result
else:
print(save_dir, ' not exists')
result_mean, result_std = average_seed(all_seed_dict)
result_data_method[f'{source_domain}-{target_domain}'] = result_mean
all_columns_mean.append([dataset, NN[STETTING][dataset][0], NN[STETTING][dataset][1], method, source_domain, target_domain] + [result_mean[term] for term in RESULTS_TERMS])
all_columns_std.append([dataset, NN[STETTING][dataset][0], NN[STETTING][dataset][1], method, source_domain, target_domain] + [result_std[term] for term in RESULTS_TERMS])
mean_data_method, _ = average_seed(result_data_method)
overall_mean.append([dataset, NN[STETTING][dataset][0], NN[STETTING][dataset][1], method] + [mean_data_method[term] for term in RESULTS_TERMS])
overall_mean_steps.append([dataset, NN[STETTING][dataset][0], NN[STETTING][dataset][1], method, step] + [mean_data_method[term] for term in RESULTS_TERMS])
all_headers = ['dataset', 'n_share', 'n_source_private', 'method', 'source', 'target'] + RESULTS_TERMS
result_path = f'{args.backbone}-{args.optimizer}-{args.base_lr}-{args.classifier_head}-{args.fixed_backbone}-{args.fixed_BN}-{args.image_augmentation}-{args.batch_size}'.replace('/','')
save_all_csv(all_headers, all_columns_mean, os.path.join(save_result_dir, STETTING, f'{step}', result_path, 'mean.csv'))
save_all_csv(all_headers, all_columns_std, os.path.join(save_result_dir, STETTING, f'{step}', result_path, 'std.csv'))
overall_headers = ['dataset', 'n_share', 'n_source_private', 'method'] + RESULTS_TERMS
save_all_csv(overall_headers, overall_mean, os.path.join(save_result_dir, STETTING, f'{step}', result_path, 'mean_average.csv'))
overall_headers_steps = ['dataset', 'n_share', 'n_source_private', 'method', 'step'] + RESULTS_TERMS
save_all_csv(overall_headers_steps, overall_mean_steps, os.path.join(save_result_dir, STETTING, result_path, 'mean_average.csv'))
def average_seed(all_seed_dict):
result_mean = {}
result_std = {}
for term in RESULTS_TERMS:
result = []
for seed in all_seed_dict.keys():
if term in all_seed_dict[seed].keys():
result.append(all_seed_dict[seed][term])
elif term == 'Closed-set OA':
result.append(all_seed_dict[seed]['Closed Set Accuracy']['OA'])
elif term == 'Closed-set AA':
result.append(all_seed_dict[seed]['Closed Set Accuracy']['AA'])
else:
result.append(all_seed_dict[seed]['OSR Accuracy'][term])
result_mean[term] = round(np.mean(result), 2)
result_std[term] = round(np.std(result), 2)
return result_mean, result_std
def save_all_csv(all_headers, all_columns, result_path):
result_dir = os.path.dirname(result_path)
if not os.path.exists(result_dir):
os.makedirs(result_dir)
with open(result_path, 'w+') as f:
writer = csv.writer(f)
writer.writerow(all_headers)
writer.writerows(all_columns)
if __name__ == "__main__":
###########################
# Directory Config (modify if using your own paths)
###########################
parser.add_argument(
"--data_dir",
type=str,
default=default.DATA_DIR,
help="where the dataset is saved",
)
parser.add_argument(
"--feature_dir",
type=str,
default=default.FEATURE_DIR,
help="where to save pre-extracted features",
)
parser.add_argument(
"--result_dir",
type=str,
default=default.RESULT_DIR,
help="where to save experiment results",
)
parser.add_argument(
"--suffix",
type=str,
default=None,
help="the save_file suffix after result_dir",
)
parser.add_argument(
"--backbone",
type=str,
default="ViT-L/14@336px",
choices=backbone_names,
help="specify the encoder-backbone to use",
)
parser.add_argument(
"--classifier_head",
type=str,
default="prototype",
choices=head_names,
help="classifier head architecture",
)
parser.add_argument(
"--fixed_backbone",
action="store_true",
help="wheather fixed backbone during training",
)
parser.add_argument(
"--fixed_BN",
action="store_true",
help="wheather fixed batch normalization layers during training",
)
parser.add_argument(
"--batch_size",
type=int,
default=32,
help="batch size for test (feature extraction and evaluation)",
)
parser.add_argument(
"--image_augmentation",
type=str,
default='none',
choices=['none', # only a single center crop
'flip', # add random flip view
'randomcrop', # add random crop view
],
help="specify the image augmentation to use.",
)
parser.add_argument(
"--optimizer",
type=str,
default="sgd",
choices=["adamw", "sgd"],
help="optimizer"
)
parser.add_argument(
'--base_lr',
type=float,
default=1e-2,
help='base learning rate'
)
args = parser.parse_args()
main(args)