-
Notifications
You must be signed in to change notification settings - Fork 0
/
wandb_fast_reid.yaml
197 lines (167 loc) · 3.32 KB
/
wandb_fast_reid.yaml
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
command:
- "srun"
- "-c"
- "8"
- "--gres=gpu:4"
- "--mem"
- "128G"
- "--time=4:00:00"
- bash
- ${program}
- "--config-file=./fastreid_configs/Market1501/bagtricks_R50.yml"
- "--num-gpus=4"
- ${args_no_hyphens}
program: wandb_fast_reid.sh
method: bayes
metric:
name: mINP
goal: maximize
parameters:
SOLVER.MAX_EPOCH:
value: 1440
TEST.EVAL_PERIOD:
value: 144
SOLVER.CHECKPOINT_PERIOD:
value: 144
INPUT.SIZE_TRAIN:
value: "(64, 64)"
INPUT.SIZE_TEST:
value: "(64, 64)"
DATALOADER.NUM_INSTANCE:
value: 4
DATALOADER.NUM_WORKERS:
value: 8
SOLVER.IMS_PER_BATCH:
value: 64
INPUT.AUGMIX.ENABLED:
value: True
INPUT.AUGMIX.PROB:
distribution: uniform
min: 0.0
max: 1
INPUT.CJ.ENABLED:
value: True
INPUT.CJ.PROB:
distribution: uniform
min: 0.0
max: 1
INPUT.CJ.BRIGHTNESS:
distribution: uniform
min: 0.0
max: 2
INPUT.CJ.CONTRAST:
distribution: uniform
min: 0.0
max: 2
INPUT.CJ.HUE:
distribution: uniform
min: 0.0
max: 0.5
INPUT.CJ.SATURATION:
distribution: uniform
min: 0.0
max: 2
INPUT.FLIP.ENABLED:
values: [True, False]
INPUT.FLIP.PROB:
value: 0.5
INPUT.ROTATION.ENABLED:
value: True
INPUT.ROTATION.DEGREES:
distribution: int_uniform
min: 60
max: 180
MODEL.LOSSES.CE.ALPHA:
value: 0.2
MODEL.LOSSES.CE.EPSILON:
value: 0.1
MODEL.LOSSES.CE.SCALE:
value: 1.0
MODEL.LOSSES.TRI.HARD_MINING:
value: True
MODEL.LOSSES.TRI.NORM_FEAT:
value: False
MODEL.LOSSES.TRI.MARGIN:
distribution: uniform
min: 0.0
max: 0.3
MODEL.LOSSES.TRI.SCALE:
value: 1.0
MODEL.QUEUE_SIZE:
value: 8192
SOLVER.BASE_LR:
distribution: uniform
max: 0.005
min: 0.00001
SOLVER.BIAS_LR_FACTOR:
value: 1.0
SOLVER.CLIP_GRADIENTS.ENABLED:
values: [False, True]
SOLVER.CLIP_GRADIENTS.CLIP_TYPE:
value: "norm"
SOLVER.CLIP_GRADIENTS.CLIP_VALUE:
value: 5.0
SOLVER.CLIP_GRADIENTS.NORM_TYPE:
value: 2.0
SOLVER.DELAY_EPOCHS:
value: 720
SOLVER.FREEZE_ITERS:
value: 1000
SOLVER.GAMMA:
value: 0.1
SOLVER.HEADS_LR_FACTOR:
value: 1.0
SOLVER.MOMENTUM:
distribution: uniform
max: 0.95
min: 0.9
SOLVER.OPT:
values: ["Adam", "SGD"]
SOLVER.SCHED:
value: "CosineAnnealingLR"
SOLVER.ETA_MIN_LR:
value: 0.0000007
SOLVER.WARMUP_FACTOR:
value: 0.1
SOLVER.WARMUP_ITERS:
value: 2000
SOLVER.WARMUP_METHOD:
value: "linear"
SOLVER.WEIGHT_DECAY:
value: 0.0005
SOLVER.WEIGHT_DECAY_BIAS:
value: 0.0005
SOLVER.WEIGHT_DECAY_NORM:
value: 0.0005
TEST.IMS_PER_BATCH:
value: 128
DATALOADER.SAMPLER_TRAIN:
value: NaiveIdentitySampler
OUTPUT_DIR:
value: runs/apparence/train
INPUT.CROP.ENABLED:
value: False
INPUT.PADDING.ENABLED:
value: False
INPUT.REA.ENABLED:
value: False
INPUT.RPT.ENABLED:
value: False
INPUT.AFFINE.ENABLED:
value: False
INPUT.AUTOAUG.ENABLED:
value: False
SOLVER.AMP.ENABLED:
value: False
TEST.FLIP.ENABLED:
value: False
TEST.ROC.ENABLED:
value: False
SOLVER.NESTEROV:
value: False
MODEL.HEADS.NUM_CLASSES:
value: 93
MODEL.PIXEL_MEAN:
value: "(0.00386753, 0.00386753, 0.00386753)"
MODEL.PIXEL_STD:
value: "(0.04723771, 0.04723771, 0.04723771)"