Hugging Face's logo

Spaces:
bkhmsi
/
Partial-Arabic-Diacritization
Sleeping

App Files Community
Partial-Arabic-Diacritization
File size: 2,222 Bytes 
d36d50b
 
d7c4b94
d36d50b
 
 
 
 
d7c4b94
d36d50b
 
 
 
 
d7c4b94
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d36d50b
 
 
 
 
 
d7c4b94
d36d50b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d7c4b94
d36d50b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
run-title: tashkeela-d2
debug: false
model-name: TD2

paths:
  base: ./dataset/ashaar
  save: ./models
  load: tashkeela-d2.pt
  load-td2: td2/tashkeela-ashaar-td2.pt
  resume: ./models/Tashkeela-D2/tashkeela-d2.pt
  constants: ./dataset/helpers/constants
  word-embs: vocab.vec
  test: test

modeling:
  "checkpoint": munael/Partial-Arabic-Diacritization-TD2
  "base_model": CAMeL-Lab/bert-base-arabic-camelbert-mix-ner
  # "base_model": UBC-NLP/MARBERTv2
  # "base_model": UBC-NLP/ARBERTv2
  "deep-cls": true
  "full-finetune": true #< From true
  "keep-token-model-layers": 2
  # "num-finetune-last-layers": 2 #
  "num-chars": 40
  "char-embed-dim": 128
  "token_hidden_size": 768
  "deep-down-proj": true
  "dropout": 0.2
  "sentence_dropout": 0.1
  "diac_model_config": {
    "vocab_size": 1,
    "num_hidden_layers": 2,
    "hidden_size": 768,
    "intermediate_size": 2304,
    "num_attention_heads": 8,
  }

loader:
  wembs-limit: -1
  num-workers: 0

train:
  epochs: 1000
  batch-size: 1
  char-embed-dim: 32
  resume: false
  resume-lr: false

  max-word-len: 13
  max-sent-len: 10

  rnn-cell: lstm
  sent-lstm-layers: 2
  word-lstm-layers: 2

  sent-lstm-units: 256
  word-lstm-units: 512
  decoder-units: 256

  sent-dropout: 0.2
  diac-dropout: 0
  final-dropout: 0.2

  sent-mask-zero: false

  lr-factor: 0.5
  lr-patience: 1
  lr-min: 1.e-7
  lr-init: 0.002

  weight-decay: 0
  vertical-dropout: 0.25
  recurrent-dropout: 0.25

  stopping-delta: 1.e-7
  stopping-patience: 3

predictor:
  batch-size: 1
  stride: 2
  window: 20
  gt-signal-prob: 0
  seed-idx: 0

sentence-break:
  stride: 2
  window: 10
  min-window: 1
  export-map: false
  files:
    - train/train.txt
    - val/val.txt
  delimeters:
    - ،
    - ؛
    - ','
    - ;
    - «
    - »
    - '{'
    - '}'
    - '('
    - ')'
    - '['
    - ']'
    - '.'
    - '*'
    - '-'
    - ':'
    - '?'
    - '!'
    - ؟


segment:
  stride: 2
  window: 10
  min-window: 1
  export-map: false
  files:
    - train/train.txt
    - val/val.txt
  delimeters:
    - ،
    - ؛
    - ','
    - ;
    - «
    - »
    - '{'
    - '}'
    - '('
    - ')'
    - '['
    - ']'
    - '.'
    - '*'
    - '-'
    - ':'
    - '?'
    - '!'
    - ؟