-
Notifications
You must be signed in to change notification settings - Fork 0
/
base_options.py
272 lines (204 loc) · 9.88 KB
/
base_options.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
#!/usr/bin/env python
# Copyright (c) Facebook, Inc. and its affiliates.
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.
import argparse
import os
# from util import util
import torch
class BaseOptions():
def __init__(self):
self.parser = argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter)
self.initialized = False
def initialize(self):
self.parser.add_argument(
"--audio_dir", type=str, default='data/feats/vggish/', help="audio dir")
self.parser.add_argument(
"--video_dir", type=str, default='data/feats/res152/',
help="video dir")
self.parser.add_argument(
"--st_dir", type=str, default='data/feats/r2plus1d_18/',
help="video dir")
self.parser.add_argument(
"--label_train", type=str, default="data/AVVP_train.csv", help="weak train csv file")
self.parser.add_argument(
"--label_val", type=str, default="data/AVVP_val_pd.csv", help="weak val csv file")
self.parser.add_argument(
"--label_test", type=str, default="data/AVVP_test_pd.csv", help="weak test csv file")
self.parser.add_argument('--batch_size', type=int, default=16, metavar='N',
help='input batch size for training (default: 16)')
self.parser.add_argument('--epochs', type=int, default=40, metavar='N',
help='number of epochs to train (default: 60)')
############# yb param ###########
self.parser.add_argument('--lr', type=float, default=3e-4, metavar='LR',
help='learning rate (default: 3e-4)')
self.parser.add_argument('--lr_mlp', type=float, default=1e-4, metavar='LR',
help='learning rate (default: 3e-4)')
self.parser.add_argument('--lr_v', type=float, default=3e-4, metavar='LR',
help='learning rate (default: 3e-4)')
self.parser.add_argument('--occ_dim', type=int, default=64, metavar='LR',
help='learning rate (default: 3e-4)')
self.parser.add_argument('--init_epoch', type=int, default=5, metavar='LR',
help='learning rate (default: 3e-4)')
############# yb param ###########
self.parser.add_argument(
"--model", type=str, default='MMIL_Net', help="with model to use")
self.parser.add_argument(
"--mode", type=str, default='train', help="with mode to use")
self.parser.add_argument('--seed', type=int, default=1, metavar='S',
help='random seed (default: 1)')
self.parser.add_argument('--log-interval', type=int, default=50, metavar='N',
help='how many batches to wait before logging training status')
self.parser.add_argument(
"--model_save_dir", type=str, default='models/', help="model save dir")
self.parser.add_argument(
"--output_dir", type=str, default='outputs/', help="model save dir")
self.parser.add_argument('--fps', type=int, default=1)
self.parser.add_argument(
"--checkpoint", type=str, default='cvpr_best',
help="save model name")
self.parser.add_argument(
'--gpu', type=str, default='0,1,2,3,4,5,6,7', help='gpu device number')
self.parser.add_argument(
'--wandb', type=int, default='0', help='weight and bias setup')
self.parser.add_argument(
'--is_v_ori', type=int, default='0', help='original visual features')
self.parser.add_argument(
'--is_a_ori', type=int, default='0', help='original audio features')
self.parser.add_argument(
'--tsne', type=int, default='0', help='run tsne or not')
self.parser.add_argument(
'--early_stop', type=int, default='5', help='weight and bias setup')
self.parser.add_argument(
'--threshold', type=float, default=0.5, help='weight and bias setup')
self.parser.add_argument(
'--save_model', action="store_true"
)
self.parser.add_argument(
'--pretrained', action="store_true"
)
self.parser.add_argument(
"--tmp", type=float, default=0.5,
help="video dir")
self.parser.add_argument(
"--noisy_label", action="store_true", default=False, help="audio dir")
self.parser.add_argument(
"--smooth", type=float, default=1,
help="video dir")
### yb param ##
self.parser.add_argument(
'--margin1', type=float, default=0.05, help='weight and bias setup')
self.parser.add_argument(
'--alpha', type=float, default=1, help='weight and bias setup')
self.parser.add_argument(
'--beta', type=float, default=1, help='weight and bias setup')
self.parser.add_argument(
'--delta', type=float, default=1, help='weight and bias setup')
self.parser.add_argument(
'--gamma', type=float, default=1, help='weight and bias setup')
self.parser.add_argument(
'--decay', type=float, default=0.1, help='decay rate')
self.parser.add_argument(
'--decay_epoch', type=float, default=10, help='decay rate')
self.parser.add_argument(
'--aug_type', type=str, default='vision', help='weight and bias setup')
self.parser.add_argument(
'--pos_pool', type=str, default='max', help='weight and bias setup')
self.parser.add_argument(
'--neg_pool', type=str, default='mean', help='weight and bias setup')
self.parser.add_argument(
'--exp', type=int, default=0, help='weight and bias setup')
self.parser.add_argument(
'--ybloss', type=int, default=1, help='decay rate')
### for transformer ###
self.parser.add_argument(
'--num_layer', type=int, default=1, help='num layer')
self.parser.add_argument(
'--num_head', type=int, default=1, help='num layer')
self.parser.add_argument(
'--prob_drop', type=float, default=0.1, help='drop out')
self.parser.add_argument(
'--prob_drop_occ', type=float, default=0.1, help='drop out')
self.parser.add_argument(
'--forward_dim', type=int, default=512, help='drop out')
self.parser.add_argument(
'--epoch_remove', type=int, default=1, help='weight and bias setup')
#######################
self.parser.add_argument(
'--audio_enc', type=int, default= 0, help='weight and bias setup')
self.parser.add_argument(
'--num_remove', type=int, default= 4, help='num of instances removing')
### for AV-ada ###
self.parser.add_argument('--audio_folder', type=str, default="", help="raw audio path")
self.parser.add_argument('--video_folder', type=str, default="", help="video frame path")
self.parser.add_argument('--audio_length', type=float, default= 1, help='audio length')
self.parser.add_argument('--num_workers', type=int, default= 0, help='worker for dataloader')
self.parser.add_argument('--model_name', type=str, default=None, help="for log")
self.parser.add_argument('--qkv_fusion', type=int, default=1, help="qkv fusion")
self.parser.add_argument('--adapter_kind', type=str, default='bottleneck', help="for log")
self.parser.add_argument('--start_tune_layers', type=int, default=0, help="tune top k")
self.parser.add_argument('--start_fusion_layers', type=int, default=0, help="tune top k")
self.parser.add_argument('--Adapter_downsample', type=int, default=16, help="tune top k")
self.parser.add_argument('--num_conv_group', type=int, default=2, help="group conv")
self.parser.add_argument('--log_path', type=str, default='', help="for log")
self.parser.add_argument('--is_audio_adapter_p1', type=int, default=0, help="TF audio adapter")
self.parser.add_argument('--is_audio_adapter_p2', type=int, default=0, help="TF audio adapter")
self.parser.add_argument('--is_audio_adapter_p3', type=int, default=0, help="TF audio adapter")
self.parser.add_argument('--is_bn', type=int, default=0, help="TF audio adapter")
self.parser.add_argument('--is_gate', type=int, default=0, help="TF audio adapter")
self.parser.add_argument('--is_multimodal', type=int, default=1, help="TF audio adapter")
self.parser.add_argument('--is_before_layernorm', type=int, default=1, help="TF audio adapter")
self.parser.add_argument('--is_post_layernorm', type=int, default=1, help="TF audio adapter")
self.parser.add_argument('--is_vit_ln', type=int, default=0, help="TF Vit")
self.parser.add_argument('--is_fusion_before', type=int, default=0, help="TF Vit")
self.parser.add_argument('--num_tokens', type=int, default=32, help="num of MBT tokens")
self.parser.add_argument('--vis_encoder_type', type=str, default="swin", help="type of visual backbone")
self.parser.add_argument('--vit_type', type=str, default=None, help="type of transformer backbone")
self.parser.add_argument('--exp_name', type=str, default="", help="name of the experiment")
self.parser.add_argument('--n_vis_tokens', type=int, default=5, help="num of MBT tokens")
self.parser.add_argument('--n_audio_tokens', type=int, default=5, help="num of MBT tokens")
self.parser.add_argument('--n_shared_tokens', type=int, default=5, help="num of MBT tokens")
self.parser.add_argument('--checkpoint_path', type=str, default='', help="for log")
self.parser.add_argument(
'--load_model', action="store_true"
)
self.parser.add_argument(
'--r', type=float, default=0.8
)
self.parser.add_argument(
'--loss2', action="store_true"
)
self.parser.add_argument(
'--freeze', action="store_true"
)
def parse(self):
if not self.initialized:
self.initialize()
self.opt = self.parser.parse_args()
str_ids = self.opt.gpu.split(',')
self.opt.gpu = []
for str_id in str_ids:
id = int(str_id)
if id >= 0:
self.opt.gpu.append(id)
# # set gpu ids
# if len(self.opt.gpu_ids) > 0:
# torch.cuda.set_device(self.opt.gpu_ids[0])
#I should process the opt here, like gpu ids, etc.
args = vars(self.opt)
print('------------ Options -------------')
for k, v in sorted(args.items()):
print('%s: %s' % (str(k), str(v)))
print('-------------- End ----------------')
# save to the disk
# expr_dir = os.path.join(self.opt.checkpoints_dir, self.opt.name)
# util.mkdirs(expr_dir)
# file_name = os.path.join(expr_dir, 'opt.txt')
# with open(file_name, 'wt') as opt_file:
# opt_file.write('------------ Options -------------\n')
# for k, v in sorted(args.items()):
# opt_file.write('%s: %s\n' % (str(k), str(v)))
# opt_file.write('-------------- End ----------------\n')
return self.opt