-
Notifications
You must be signed in to change notification settings - Fork 19
/
args.py
357 lines (344 loc) · 10.3 KB
/
args.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
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
import argparse
import os
PRESAVE_DIR = "TOFILL"
MODEL_DIR = "TOFILL"
DATA_DIR = "TOFILL"
SSD_DIR = "TOFILL"
NLTK_FOLDER = "TOFILL"
name2folder = {
"youcook": "YouCook2",
"htm": "howto100m",
"chapters": "AllChapters",
"vitt": "ViTT"
}
def get_args_parser():
parser = argparse.ArgumentParser("Set Vid2Seq", add_help=False)
# Dataset specific
parser.add_argument(
"--combine_datasets",
nargs="+",
help="list of datasets to combine for training",
required=True,
)
parser.add_argument(
"--combine_datasets_val",
nargs="+",
help="list of datasets to combine for eval",
default=[],
)
parser.add_argument(
"--howto100m_train_csv_path",
default=os.path.join(DATA_DIR, name2folder["htm"], "htm_vid2seq.csv"),
)
parser.add_argument(
"--howto100m_features_path",
default=os.path.join(SSD_DIR, "howto100m_clip_features"),
)
parser.add_argument(
"--howto100m_subtitles_path",
default=os.path.join(SSD_DIR, "htm_sentences"),
)
parser.add_argument(
"--youcook_features_path",
default=os.path.join(DATA_DIR, name2folder["youcook"], "clipvitl14.pth"),
)
parser.add_argument(
"--youcook_train_json_path",
default=os.path.join(DATA_DIR, name2folder["youcook"], "train.json"),
)
parser.add_argument(
"--youcook_val_json_path",
default=os.path.join(DATA_DIR, name2folder["youcook"], "val.json"),
)
parser.add_argument(
"--youcook_subtitles_path",
default=os.path.join(DATA_DIR, name2folder["youcook"], "youcook2_asr_align_proc.pkl"),
)
parser.add_argument(
"--vitt_features_path",
default=os.path.join(DATA_DIR, name2folder["vitt"], "clipvitl14.pth"),
)
parser.add_argument(
"--vitt_train_json_path",
default=os.path.join(DATA_DIR, name2folder["vitt"], "train.json"),
)
parser.add_argument(
"--vitt_val_json_path",
default=os.path.join(DATA_DIR, name2folder["vitt"], "dev.json"),
)
parser.add_argument(
"--vitt_test_json_path",
default=os.path.join(DATA_DIR, name2folder["vitt"], "test.json"),
)
parser.add_argument(
"--vitt_subtitles_path",
default=os.path.join(DATA_DIR, name2folder["vitt"], "subtitles_align_proc.pkl"),
)
parser.add_argument(
"--chapters_features_path",
default=os.path.join(SSD_DIR, "chapters_clipvitl14_features"),
)
parser.add_argument(
"--chapters_train_json_path",
default=os.path.join(DATA_DIR, name2folder["chapters"], "chapters_dvc_train.json"),
)
parser.add_argument(
"--chapters_val_json_path",
default=os.path.join(DATA_DIR, name2folder["chapters"], "chapters_dvc_val.json"),
)
parser.add_argument(
"--chapters_test_json_path",
default=os.path.join(DATA_DIR, name2folder["chapters"], "chapters_dvc_test.json"),
)
parser.add_argument(
"--chapters_subtitles_path",
default=os.path.join(SSD_DIR, "allchapters_asr"),
)
# Training hyper-parameters
parser.add_argument(
"--denoising", default=1., type=float, help="denoising loss coef"
)
parser.add_argument(
"--generative", default=1., type=float, help="generative loss coef"
)
parser.add_argument("--genasr", action="store_true", help="baseline that generates asr and not chapters")
parser.add_argument("--random", action="store_true", help="random baseline")
parser.add_argument(
"--mask_prob",
type=float,
default=0.25,
help="masking probability for the denoising objective",
)
parser.add_argument(
"--mask_len",
type=int,
default=5,
help="masking average span length for the denoising objective",
)
parser.add_argument("--lr", default=3e-4, type=float, help="learning rate")
parser.add_argument(
"--beta1", default=0.9, type=float, help="Adam optimizer parameter"
)
parser.add_argument(
"--beta2", default=0.999, type=float, help="Adam optimizer parameter"
)
parser.add_argument(
"--batch_size", default=2, type=int, help="batch size used for training"
)
parser.add_argument(
"--batch_size_val",
default=2,
type=int,
help="batch size used for eval",
)
parser.add_argument("--weight_decay", default=0, type=float)
parser.add_argument(
"--epochs", default=20, type=int, help="number of training epochs"
)
parser.add_argument("--optimizer", default="adam", type=str)
parser.add_argument(
"--label_smoothing", default=0.1, type=float, help="label smoothing"
)
parser.add_argument(
"--clip_max_norm", default=1., type=float, help="gradient clipping max norm"
)
parser.add_argument(
"--schedule",
default="",
choices=["", "cosine_with_warmup"],
help="learning rate decay schedule, default is constant",
)
parser.add_argument(
"--fraction_warmup_steps",
default=0.1,
type=float,
help="fraction of number of steps used for warmup when using cosine schedule",
)
parser.add_argument(
"--eval_skip",
default=1,
type=int,
help='do evaluation every "eval_skip" epochs',
)
parser.add_argument(
"--print_freq",
type=int,
default=100,
help="print log every print_freq iterations",
)
# Run specific
parser.add_argument(
"--save_dir", default="", help="path where to save, empty for no saving"
)
parser.add_argument(
"--presave_dir",
default=PRESAVE_DIR,
help="the actual save_dir is an union of presave_dir and save_dir",
)
parser.add_argument("--device", default="cuda", help="device to use")
parser.add_argument("--seed", default=42, type=int, help="random seed")
parser.add_argument(
"--load",
default="",
help="path to load checkpoint",
)
parser.add_argument(
"--resume",
action="store_true",
help="continue training if loading checkpoint",
)
parser.add_argument(
"--start-epoch", default=0, type=int, metavar="N", help="start epoch"
)
parser.add_argument("--eval", action="store_true", help="only run evaluation")
parser.add_argument(
"--num_workers", default=3, type=int, help="number of workers for dataloader"
)
# Distributed training parameters
parser.add_argument(
"--world-size", default=1, type=int, help="number of distributed processes"
)
parser.add_argument(
"--dist-url", default="env://", help="url used to set up distributed training"
)
# Model parameters
parser.add_argument(
"--model_name",
default="t5-base",
choices=(
"t5-base", os.path.join(MODEL_DIR, "7BHF"), "Salesforce/blip2-flan-t5-xl"
),
)
parser.add_argument(
"--bert_name",
default="bert-base-uncased",
choices=(
"bert-base-uncased"
),
)
parser.add_argument(
"--text_encoder_dropout", default=0.1, type=float, help="dropout to use in the text encoder"
)
parser.add_argument(
"--text_decoder_dropout", default=0.1, type=float, help="dropout to use in the text decoder"
)
parser.add_argument(
"--visual_encoder_dropout", default=0.1, type=float, help="dropout to use in the visual encoder"
)
parser.add_argument(
"--max_feats",
type=int,
default=100,
help="maximum number of video features considered, one per frame",
)
parser.add_argument(
"--features_dim",
type=int,
default=768,
help="dimension of the visual embedding space",
)
parser.add_argument(
"--embedding_dim",
type=int,
default=768,
help="dimension of the language modeling space",
)
parser.add_argument(
"--mlp_dim",
type=int,
default=2048,
help="dimension of the visual encoder mlp",
)
parser.add_argument(
"--depth",
type=int,
default=12,
help="number of layers of visual encoder",
)
parser.add_argument(
"--heads",
type=int,
default=12,
help="number of heads of visual encoder",
)
parser.add_argument(
"--num_bins",
type=int,
default=100,
help="number of quantization bins for the time tokens",
)
parser.add_argument(
"--no_video",
dest="use_video",
action="store_false",
help="disables usage of video",
)
parser.add_argument(
"--no_speech",
dest="use_speech",
action="store_false",
help="disables usage of speech",
)
parser.add_argument(
"--max_input_tokens",
type=int,
default=1000,
help="maximum number of tokens in the input speech",
)
parser.add_argument(
"--max_output_tokens",
type=int,
default=256,
help="maximum number of tokens in the output sequence of dense captions",
)
parser.add_argument(
"--num_beams",
type=int,
default=4,
help="beam search size",
)
parser.add_argument(
"--length_penalty",
type=float,
default=1.,
help="length penalty for beam search",
)
parser.add_argument(
"--repetition_penalty",
type=float,
default=1.,
help="repetition penalty for beam search",
)
parser.add_argument(
"--top_p",
type=float,
default=0.9,
help="nucleus sampling parameter",
)
# BLIP-2 Model parameters
parser.add_argument(
"--blip2_model_name",
default="pretrain_flant5xl_vitL",
choices=(
"pretrain_flant5xl_vitL"
),
)
parser.add_argument(
"--resolution",
type=int,
default=224,
help="spatial resolution for raw video",
)
parser.add_argument(
"--video_example",
default="",
type=str,
help="path to a video example for demo",
)
parser.add_argument(
"--asr_example",
default="",
type=str,
help="path to a ASR example for demo",
)
return parser