-
Notifications
You must be signed in to change notification settings - Fork 35
/
opts.py
125 lines (111 loc) · 7.77 KB
/
opts.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
import argparse
import os
def parse_opt():
parser = argparse.ArgumentParser()
parser.add_argument('--debug', type=bool, default=False)
parser.add_argument('--option', type=str, default='train', help='train | test')
parser.add_argument('--id', type=str, default='default', help='an id identifying this run/job')
# Data input settings
parser.add_argument('--data_dir', type=str, default='/mnt/sshd/xwang/VIST')
# parser.add_argument('--data_dir', type=str, default='../datasets/VIST')
parser.add_argument('--desc_h5', type=str, default='VIST/description.h5')
parser.add_argument('--story_h5', type=str, default='VIST/story.h5')
parser.add_argument('--full_story_h5', type=str, default='VIST/full_story.h5')
parser.add_argument('--story_line_json', type=str, default='VIST/story_line.json')
parser.add_argument('--resume_from', type=str, default=None,
help="""continue training from saved model at this path. Path must contain files saved by previous training process:
'infos.pkl' : configuration;
'checkpoint' : paths to model file(s) (created by tf).
Note: this file contains absolute paths, be careful when moving files around;
'model.ckpt-*' : file(s) with model definition (created by tf)""")
parser.add_argument('--start_from_model', type=str, default=None)
# Model settings
parser.add_argument('--model', type=str, default="BaseModel")
parser.add_argument('--task', type=str, default="story_telling", help='story_telling')
parser.add_argument('--rnn_type', type=str, default='gru', help='gru, lstm')
parser.add_argument('--hidden_dim', type=int, default=512)
parser.add_argument('--num_layers', type=int, default=1, help='number of layers in the RNN')
parser.add_argument('--word_embed_dim', type=int, default=512,
help='the encoding size of each token in the vocabulary, and the image.')
parser.add_argument('--feat_size', type=int, default=2048, help='2048 for resnet, 4096 for vgg')
parser.add_argument('--conv_feat_size', type=int, default=7 * 7 * 2048, help='2048 for resnet, 4096 for vgg')
parser.add_argument('--use_conv', type=bool, default=False)
parser.add_argument('--story_size', type=int, default=5,
help='number of images/sentences appearing in each story')
parser.add_argument('--with_position', type=bool, default=False,
help='whether to use position embedding for the image feature')
# Optimization: General
parser.add_argument('--max_epochs', type=int, default=100,
help='number of epochs')
parser.add_argument('--shuffle', type=bool, default=True,
help='set to True to have the data reshuffled at every epoch during training ')
parser.add_argument('--batch_size', type=int, default=64, help='minibatch size')
parser.add_argument('--workers', type=int, default=8,
help='number of workers to load data')
parser.add_argument('--grad_clip', type=float, default=10,
help='clip gradients at this value')
parser.add_argument('--visual_dropout', type=float, default=0.2,
help='strength of dropout in the Language Model RNN')
parser.add_argument('--dropout', type=float, default=0.5,
help='strength of dropout in the Language Model RNN')
parser.add_argument('--beam_size', type=int, default=1,
help='indicates number of beams in beam search. This is only used in the evaluation mode')
# Optimization: for the Language Model
parser.add_argument('--optim', type=str, default='Adam',
help='RMSprop|SGD|momSGD|Adam|Adadelta|YF')
parser.add_argument('--learning_rate', type=float, default=4e-4, help='learning rate')
parser.add_argument('--learning_rate_decay_start', type=int, default=-1,
help='from which epoch to start decaying learning rate? (-1 = dont)')
parser.add_argument('--learning_rate_decay_every', type=int, default=5,
help='every how many epochs thereafter to drop LR')
parser.add_argument('--learning_rate_decay_rate', type=float, default=0.5,
help='every how many epochs thereafter to drop LR')
parser.add_argument('--momentum', type=float, default=0.9,
help='momentum, only used in rmsprop & sgd')
parser.add_argument('--weight_decay', type=float, default=0, help='weight_decay')
# Adam optimizer setting
parser.add_argument('--optim_alpha', type=float, default=0.8, help='alpha for momentum')
parser.add_argument('--optim_beta', type=float, default=0.999, help='beta for adam')
parser.add_argument('--optim_epsilon', type=float, default=1e-8,
help='epsilon that goes into denominator for smoothing')
# Reinforcement learning
parser.add_argument('--start_rl', type=int, default=-1,
help='at what epoch to start RL training, -1 means never')
parser.add_argument('--reward_type', type=str, default='METEOR',
help="CIDEr | ROUGE_L | METEOR | Bleu_4 | Bleu_3")
parser.add_argument('--rl_weight', type=float, default=0.5)
parser.add_argument('--cached_tokens', type=str, default='VIST/VIST-train-words',
help='used to compute CIDEr reward')
parser.add_argument('--use_feature_with_context', type=bool, default=False)
parser.add_argument('--decoding_method_DISC', type=str, default='sample', help='greedy | sample')
# Schedule sampling
parser.add_argument('--scheduled_sampling_start', type=int, default=0,
help='at what epoch to start decay gt probability, -1 means never')
parser.add_argument('--scheduled_sampling_increase_every', type=int, default=5,
help='every how many epochs to increase scheduled sampling probability')
parser.add_argument('--scheduled_sampling_increase_prob', type=float, default=0.05,
help='How much to update the prob')
parser.add_argument('--scheduled_sampling_max_prob', type=float, default=0.25,
help='Maximum scheduled sampling prob.')
# Evaluation/Checkpointing
parser.add_argument('--metric', type=str, default='METEOR',
help="XE | CIDEr | ROUGE_L | METEOR | Bleu_4 | Bleu_3")
parser.add_argument('--save_checkpoint_every', type=int, default=1000,
help='how often to save a model checkpoint (in iterations)')
parser.add_argument('--checkpoint_path', type=str, default='data/save',
help='directory to store checkpointed models')
parser.add_argument('--losses_log_every', type=int, default=10,
help='How often do we snapshot losses, for inclusion in the progress dump? (0 = disable)')
parser.add_argument('--load_best_score', type=bool, default=True,
help='whether to load previous best score when resuming training.')
parser.add_argument('--prefix', type=str, default='fc/', help="feature folder")
# misc
parser.add_argument('--always', type=str, default=None, help='always train one model, no alternating training')
parser.add_argument('--D_iter', type=int, default=100, help='Discriminator update iterations')
parser.add_argument('--G_iter', type=int, default=100, help='Generator update iterations')
parser.add_argument('--activation', type=str, default="sign",
help='the last activation function of the reward model: sign | tahn')
args = parser.parse_args()
if args.start_rl >= 0:
args.metric = args.reward_type
return args