forked from monologg/R-BERT
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbert_base_sweep.yaml
134 lines (134 loc) · 2.15 KB
/
bert_base_sweep.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
method: bayes
metric:
goal: maximize
name: eval/f1
parameters:
# adam_beta1:
# distribution: uniform
# max: 1.8
# min: 0.45
# adam_beta2:
# distribution: uniform
# max: 1.998
# min: 0.4995
# adam_epsilon:
# distribution: uniform
# max: 2e-08
# min: 5e-09
learning_rate:
values:
- 3e-05
# max_grad_norm:
# distribution: uniform
# max: 2
# min: 0.5
num_train_epochs_frozen:
values:
- 0.0
- 1.0
- 2.0
add_sep_token:
values:
- "true"
# - "false"
entity_mention:
values:
- "entity_start"
- "max_pooling",
- "max_pooling_with_markups",
- "avg_pooling",
- "avg_pooling_with_markups"
dropout_rate:
values:
- 0.15
- 0.20
- 0.25
label_smoothing_epsilon:
values:
- 0.25
- 0.30
- 0.35
fc1_d1_layer_output_size:
values:
- 20
- 30
fc1_d2_layer_output_size:
values:
- 5000
- 6000
- 7000
fc2_layer_output_size:
values:
- 5000
- 6000
- 7000
first_layer_d1:
distribution: int_uniform
max: 5
min: 0
last_layer_d1:
distribution: int_uniform
max: 10
min: 9
first_layer_d2:
distribution: int_uniform
max: 2
min: 0
second_to_last_layer_d2:
distribution: int_uniform
max: 10
min: 9
skip_1_d1:
distribution: categorical
values:
# - "true"
- "false"
skip_1_d2:
distribution: categorical
values:
# - "true"
- "false"
skip_2_d1:
distribution: categorical
values:
# - "true"
- "false"
skip_2_d2:
distribution: categorical
values:
# - "true"
- "false"
seed:
distribution: categorical
values:
- 0
# - 1
# - 2
# - 3
program: main.py
command:
- ${env}
- ${interpreter}
- ${program}
- --disable_tqdm
- true
- --do_train
- --do_eval
- --output_dir
- ./output
- --overwrite_output_dir
- --per_device_eval_batch_size
- 16
- --per_device_train_batch_size
- 8
- --evaluation_strategy
- epoch
- --model_name_or_path
- bert-base-uncased
- --run_name
- bert-base-uncased
- --save_steps
- 0
- --num_train_epochs
- 12
- ${args}