-
Notifications
You must be signed in to change notification settings - Fork 0
/
joint_config.yaml
84 lines (65 loc) · 1.37 KB
/
joint_config.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
# train config
msg_info: '15-restaurant'
random_seed: 2942435
num_epochs: 100
batch_size: 8
init_lr: 0.0012
decay_lr: 0.8
minimum_lr: 0.0002
patience: 10
grad_clipping: 10.0
unk_ratio: 0.4
weight_decay: 5e-5
# embedding config
use_sentence_vec: false
sent_vec_project: 0
sent_vec_project_activation: 'linear'
#dp_sent_vec: 0.55
# concat sentence vector after BiLSTM
# if cat_sent_after_rnn is false, we concat sent_vec and word_embed, before BiLSTM
cat_sent_after_rnn: false
encoder_project: 0 #200
# use random initialized word embedding
random_word_embed: false
# use pre-trained word embedding, not contextualized one.
use_pretrain_embed: true
# tuning pre-trained embedding
pretrain_embed_tune: true
word_embed_trainable: true
word_embed_dim: 300
word_project_dim: 200
use_char_rnn: false
char_embed_dim: 40
char_rnn_dim: 40
dp_emb: 0.43
use_pos: true
pos_embed_dim: 30
use_dep: true
dep_embed_dim: 30
treeLSTM_dim: 100
# LSTM and stack LSTM config
use_rnn_encoder: true
encoder_layer: 1
rnn_dim: 150
dp_rnn: 0.4
#grn_dim: 180
#grn_steps: 5
lmda_rnn_dim: 300
#part_ent_rnn_dim: 100
action_rnn_dim: 60
out_rnn_dim: 60
dp_state: 0.15
dp_state_h: 0.1
dp_buffer: 0.
dp_stack: 0.15
# output config
output_hidden_dim: 250
dp_out: 0.2
# for tree rnn config
action_embed_dim: 50
prd_embed_dim: 50
role_embed_dim: 50
#arg_embed_dim: 50
use_arg_type_tree: false
#l2 norm
l2_norm: 0.01