-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathconfig.py
160 lines (145 loc) · 5.86 KB
/
config.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
import os
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('--dataset', type=str, default="pang04", help='dataset?')
parser.add_argument('--outdir', type=str, default=None, help='output dir?')
parser.add_argument(
'--utility', type=str, default="submod", help='utility model?')
parser.add_argument('--k', type=int, default=None, help='num clusters?')
parser.add_argument('--seed', type=int, default=1, help='seed?')
parser.add_argument('--budget', type=int, default=500, help='budget?')
parser.add_argument(
'--cric_cls', type=int, default=1, help='Index of critical class')
parser.add_argument('--num_runs', type=int, default=10, help='num runs?')
parser.add_argument('--start_seed', type=int, default=0, help='starting seed?')
parser.add_argument(
'--regret',
action="store_true",
default=False,
help='Plot regret instead?')
parser.add_argument(
'--tao', type=float, default=0.65, help='threshold for cric class?')
parser.add_argument(
'--var', type=float, default=None, help='gaussian variance?')
parser.add_argument(
'--num_processes', type=int, default=10, help='num processes?')
parser.add_argument(
'--clip', type=int, default=50, help='Clip plot at budget?')
parser.add_argument(
'--cluster_algo', type=str, default="kmeans_both", help="clustering algo?")
parser.add_argument('--clf', type=str, default="logistic", help="classifier?")
parser.add_argument('--rep', type=str, default="svd", help="representation?")
parser.add_argument(
"--elbow",
action="store_true",
default=False,
help="run elbow method and exit")
# args = parser.parse_args()
# if args.outdir is None:
# args.outdir = "out_" + args.dataset
# if args.clip is None:
# args.clip = args.budget
# if args.dataset == "pang04":
# if args.k is None:
# args.k = 6
# if args.var is None:
# args.var = 0.001
# elif args.dataset == "pang05":
# if args.k is None:
# args.k = 6
# if args.var is None:
# args.var = 0.001
# elif args.dataset == "mcauley15":
# if args.k is None:
# args.k = 6
# if args.var is None:
# args.var = 1e-4
# elif args.dataset == "kaggle13":
# if args.k is None:
# args.k = 6
# if args.var is None:
# args.var = 1e-3
# args.clf = "kagglecnn"
# args.rep = "pixelsvd"
# print ("clusters", args.k)
# clf_config_str = "%s_%s" % (args.dataset, args.clf)
# clf_pred_file = os.path.join(args.outdir, clf_config_str + ".npz")
# clf_ckpt = os.path.join(args.outdir, "checkpoint_" + clf_config_str + ".ckpt")
# config_str = "%s_%s_%s_%s_cls_%d_%s_k_%d_b_%d_%.3f" % (args.dataset, args.clf,
# args.rep, args.utility,
# args.cric_cls,
# args.cluster_algo,
# args.k, args.budget,
# args.var)
# util_config_str = "%s_%s_%s_%s_cls_%d_%.3f" % (args.dataset, args.clf,
# args.rep, args.utility,
# args.cric_cls, args.var)
# algo_config_str = util_config_str + "_b_%d" % args.budget
# prior_types = []
# if args.utility == "submod":
# prior_types = [
# "conf",
# # "uniform",
# ]
# name_map = {
# "greedy_optimal_noupdate": "Upper bound (tractable)",
# "greedy_conf_noupdate": "Greedy-conf-fix",
# "greedy_uniform_noupdate": "Greedy-uniform",
# "uub_bandit": "UUB (Lakkaraju et al. 2017)",
# "greedy_Most uncertain": "Most-uncertain",
# "greedy_conf_cluster": "Greedy-conf",
# "greedy_uniform_cluster": "Greedy-uniform"
# }
# linestyle_map = {
# "greedy_optimal_noupdate": "--",
# "greedy_conf_noupdate": "-.",
# "greedy_uniform_noupdate": "--",
# "uub_bandit": "-.",
# "greedy_Most uncertain": ":",
# "greedy_conf_cluster": "-",
# "greedy_uniform_cluster": ":"
# }
# color_map = {
# "greedy_optimal_noupdate": "orange",
# "greedy_conf_noupdate": "black",
# "greedy_uniform_noupdate": "blue",
# "uub_bandit": "green",
# "greedy_Most uncertain": "red",
# "greedy_conf_cluster": "black",
# "greedy_uniform_cluster": "black"
# }
# cluster_algos = [
# # "kmeans"
# ]
# bandit_algos = [
# # "optimal",
# # "submodgreedy",
# # "ucb",
# # "random",
# # "epsilongreedy",
# "uub",
# ]
# utilityfile = os.path.join(args.outdir, "utility_%s.npz" % util_config_str)
# unkunk_histfile = os.path.join(args.outdir,
# "unkunkhist_%s.pdf" % util_config_str)
# unkunk_2dfile = os.path.join(args.outdir, "unkunk2d_%s.pdf" % util_config_str)
# coverage_scatterfile = os.path.join(args.outdir,
# "coverage_%s.pdf" % util_config_str)
# coverage_scatterfile2 = os.path.join(
# args.outdir, "coveragetimesconf_%s.pdf" % util_config_str)
# distance_histfile = os.path.join(
# args.outdir, "utility_%s_%s_%s_%s_cls_%d_%.3f_hist.pdf" %
# (args.dataset, args.clf, args.rep, args.utility, args.cric_cls, args.var))
# most_uncertainfile = os.path.join(args.outdir,
# algo_config_str + "_most_uncertain.npy")
# clusterfile = os.path.join(args.outdir, "cluster_%s_%s_%s_%d.npy" %
# (args.dataset, args.rep, args.cluster_algo, args.k))
# repfile = os.path.join(args.outdir, "rep_%s_%s.npy" % (args.dataset, args.rep))
# def get_algodir(algo):
# algodir = os.path.join(args.outdir, config_str + "_" + algo)
# return algodir
# def get_adap_greedy_file(prior_type, update_type, cluster_algo=""):
# algofile = os.path.join(args.outdir,
# algo_config_str + "greedy_%s_%s%s.npy" %
# (prior_type, update_type, cluster_algo))
# return algofile