Hugging Face's logo

B-K
/
umt5-thai-g2p-v2-pretraining-0.5k

Safetensors
umt5
Model card Files
xet
 Community
B-K commited on
Commit
64ae065
·
verified ·
1 Parent(s): 8b6a66e

Training in progress, epoch 1

Browse files
Files changed (6) hide show
  1. config.json +35 -0
  2. model.safetensors +3 -0
  3. special_tokens_map.json +30 -0
  4. tokenizer.json +2130 -0
  5. tokenizer_config.json +46 -0
  6. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/umt5-small",
3
+ "architectures": [
4
+ "UMT5ForConditionalGeneration"
5
+ ],
6
+ "bos_token_id": 2,
7
+ "classifier_dropout": 0.0,
8
+ "d_ff": 1024,
9
+ "d_kv": 64,
10
+ "d_model": 512,
11
+ "decoder_start_token_id": 0,
12
+ "dense_act_fn": "gelu_new",
13
+ "dropout_rate": 0.1,
14
+ "eos_token_id": 1,
15
+ "feed_forward_proj": "gated-gelu",
16
+ "initializer_factor": 1.0,
17
+ "is_encoder_decoder": true,
18
+ "is_gated_act": true,
19
+ "layer_norm_epsilon": 1e-06,
20
+ "max_new_tokens": 64,
21
+ "model_type": "umt5",
22
+ "num_decoder_layers": 8,
23
+ "num_heads": 6,
24
+ "num_layers": 8,
25
+ "pad_token_id": 0,
26
+ "relative_attention_max_distance": 128,
27
+ "relative_attention_num_buckets": 32,
28
+ "scalable_attention": true,
29
+ "tie_word_embeddings": false,
30
+ "tokenizer_class": "T5Tokenizer",
31
+ "torch_dtype": "float32",
32
+ "transformers_version": "4.47.0",
33
+ "use_cache": true,
34
+ "vocab_size": 500
35
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:86cd48d962cdf35941ab60fef54eae00660893cd22ed5f0729fbd0129236e2f1
3
+ size 178331440
special_tokens_map.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "<pad>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "unk_token": {
24
+ "content": "<unk>",
25
+ "lstrip": false,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ }
30
+ }
tokenizer.json ADDED
@@ -0,0 +1,2130 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "1.0",
3
+ "truncation": null,
4
+ "padding": null,
5
+ "added_tokens": [
6
+ {
7
+ "id": 0,
8
+ "content": "<pad>",
9
+ "single_word": false,
10
+ "lstrip": false,
11
+ "rstrip": false,
12
+ "normalized": false,
13
+ "special": true
14
+ },
15
+ {
16
+ "id": 1,
17
+ "content": "</s>",
18
+ "single_word": false,
19
+ "lstrip": false,
20
+ "rstrip": false,
21
+ "normalized": false,
22
+ "special": true
23
+ },
24
+ {
25
+ "id": 2,
26
+ "content": "<s>",
27
+ "single_word": false,
28
+ "lstrip": false,
29
+ "rstrip": false,
30
+ "normalized": false,
31
+ "special": true
32
+ },
33
+ {
34
+ "id": 3,
35
+ "content": "<unk>",
36
+ "single_word": false,
37
+ "lstrip": false,
38
+ "rstrip": false,
39
+ "normalized": false,
40
+ "special": true
41
+ }
42
+ ],
43
+ "normalizer": null,
44
+ "pre_tokenizer": null,
45
+ "post_processor": {
46
+ "type": "TemplateProcessing",
47
+ "single": [
48
+ {
49
+ "Sequence": {
50
+ "id": "A",
51
+ "type_id": 0
52
+ }
53
+ },
54
+ {
55
+ "SpecialToken": {
56
+ "id": "</s>",
57
+ "type_id": 0
58
+ }
59
+ }
60
+ ],
61
+ "pair": [
62
+ {
63
+ "Sequence": {
64
+ "id": "A",
65
+ "type_id": 0
66
+ }
67
+ },
68
+ {
69
+ "Sequence": {
70
+ "id": "B",
71
+ "type_id": 1
72
+ }
73
+ }
74
+ ],
75
+ "special_tokens": {
76
+ "</s>": {
77
+ "id": "</s>",
78
+ "ids": [
79
+ 1
80
+ ],
81
+ "tokens": [
82
+ "</s>"
83
+ ]
84
+ }
85
+ }
86
+ },
87
+ "decoder": null,
88
+ "model": {
89
+ "type": "BPE",
90
+ "dropout": null,
91
+ "unk_token": null,
92
+ "continuing_subword_prefix": null,
93
+ "end_of_word_suffix": null,
94
+ "fuse_unk": false,
95
+ "byte_fallback": false,
96
+ "ignore_merges": false,
97
+ "vocab": {
98
+ "<pad>": 0,
99
+ "</s>": 1,
100
+ "<s>": 2,
101
+ "<unk>": 3,
102
+ "'": 4,
103
+ "-": 5,
104
+ ".": 6,
105
+ "/": 7,
106
+ "1": 8,
107
+ "a": 9,
108
+ "b": 10,
109
+ "d": 11,
110
+ "e": 12,
111
+ "f": 13,
112
+ "h": 14,
113
+ "i": 15,
114
+ "j": 16,
115
+ "k": 17,
116
+ "l": 18,
117
+ "m": 19,
118
+ "n": 20,
119
+ "o": 21,
120
+ "p": 22,
121
+ "r": 23,
122
+ "s": 24,
123
+ "t": 25,
124
+ "u": 26,
125
+ "w": 27,
126
+ "ŋ": 28,
127
+ "ɔ": 29,
128
+ "ɕ": 30,
129
+ "ɛ": 31,
130
+ "ɤ": 32,
131
+ "ɯ": 33,
132
+ "ʔ": 34,
133
+ "ʰ": 35,
134
+ "ː": 36,
135
+ "˥": 37,
136
+ "˦": 38,
137
+ "˧": 39,
138
+ "˨": 40,
139
+ "˩": 41,
140
+ "̚": 42,
141
+ "̯": 43,
142
+ "͡": 44,
143
+ "ก": 45,
144
+ "ข": 46,
145
+ "ฃ": 47,
146
+ "ค": 48,
147
+ "ฅ": 49,
148
+ "ฆ": 50,
149
+ "ง": 51,
150
+ "จ": 52,
151
+ "ฉ": 53,
152
+ "ช": 54,
153
+ "ซ": 55,
154
+ "ฌ": 56,
155
+ "ญ": 57,
156
+ "ฎ": 58,
157
+ "ฏ": 59,
158
+ "ฐ": 60,
159
+ "ฑ": 61,
160
+ "ฒ": 62,
161
+ "ณ": 63,
162
+ "ด": 64,
163
+ "ต": 65,
164
+ "ถ": 66,
165
+ "ท": 67,
166
+ "ธ": 68,
167
+ "น": 69,
168
+ "บ": 70,
169
+ "ป": 71,
170
+ "ผ": 72,
171
+ "ฝ": 73,
172
+ "พ": 74,
173
+ "ฟ": 75,
174
+ "ภ": 76,
175
+ "ม": 77,
176
+ "ย": 78,
177
+ "ร": 79,
178
+ "ฤ": 80,
179
+ "ล": 81,
180
+ "ฦ": 82,
181
+ "ว": 83,
182
+ "ศ": 84,
183
+ "ษ": 85,
184
+ "ส": 86,
185
+ "ห": 87,
186
+ "ฬ": 88,
187
+ "อ": 89,
188
+ "ฮ": 90,
189
+ "ฯ": 91,
190
+ "ะ": 92,
191
+ "ั": 93,
192
+ "า": 94,
193
+ "ำ": 95,
194
+ "ิ": 96,
195
+ "ี": 97,
196
+ "ึ": 98,
197
+ "ื": 99,
198
+ "ุ": 100,
199
+ "ู": 101,
200
+ "ฺ": 102,
201
+ "฿": 103,
202
+ "เ": 104,
203
+ "แ": 105,
204
+ "โ": 106,
205
+ "ใ": 107,
206
+ "ไ": 108,
207
+ "ๅ": 109,
208
+ "็": 110,
209
+ "่": 111,
210
+ "้": 112,
211
+ "๊": 113,
212
+ "๋": 114,
213
+ "์": 115,
214
+ "ํ": 116,
215
+ "‍": 117,
216
+ "˩.": 118,
217
+ "˧.": 119,
218
+ "aː": 120,
219
+ "˦˥": 121,
220
+ "˦˥.": 122,
221
+ "˨˩.": 123,
222
+ "˥˩.": 124,
223
+ "aːn": 125,
224
+ "˩˦": 126,
225
+ "˩˩˦": 127,
226
+ "ʰa": 128,
227
+ "ɔː": 129,
228
+ "้า": 130,
229
+ "n˧.": 131,
230
+ "tɕ": 132,
231
+ "้าน": 133,
232
+ "baːn": 134,
233
+ "บ้าน": 135,
234
+ "baːn˥˩.": 136,
235
+ "˩˩˦.": 137,
236
+ "pʰ": 138,
237
+ "a˨˩.": 139,
238
+ "kʰ": 140,
239
+ "a˦˥.": 141,
240
+ "˨˩": 142,
241
+ "tʰ": 143,
242
+ "aː˧.": 144,
243
+ "n˧": 145,
244
+ "m˧.": 146,
245
+ "iː": 147,
246
+ "˥˩": 148,
247
+ "ŋ˧.": 149,
248
+ "aj": 150,
249
+ "ʰa˦˥.": 151,
250
+ "at": 152,
251
+ "ɛː": 153,
252
+ "าร": 154,
253
+ "eː": 155,
254
+ "อง": 156,
255
+ "aː˧": 157,
256
+ "oː": 158,
257
+ "i˦˥.": 159,
258
+ "tɕʰ": 160,
259
+ "uː": 161,
260
+ "it": 162,
261
+ "aːn˧.": 163,
262
+ "sa˨˩.": 164,
263
+ "pʰa": 165,
264
+ "am˧.": 166,
265
+ "ŋ˧": 167,
266
+ "on˧.": 168,
267
+ "ระ": 169,
268
+ "ัน": 170,
269
+ "การ": 171,
270
+ "ua": 172,
271
+ "tʰa": 173,
272
+ "่า": 174,
273
+ "ra˦˥.": 175,
274
+ "kaːn˧.": 176,
275
+ "an˧.": 177,
276
+ "ิน": 178,
277
+ "i˨˩.": 179,
278
+ "ŋ˩˩˦.": 180,
279
+ "j˧.": 181,
280
+ "aːt": 182,
281
+ "หน": 183,
282
+ "าม": 184,
283
+ "อย": 185,
284
+ "ีย": 186,
285
+ "iː˧.": 187,
286
+ "ia": 188,
287
+ "ɯa": 189,
288
+ "รา": 190,
289
+ "รั": 191,
290
+ "m˧": 192,
291
+ "k˨˩.": 193,
292
+ "ตำ": 194,
293
+ "tam˧.": 195,
294
+ "บล": 196,
295
+ "าน": 197,
296
+ "tʰa˦˥.": 198,
297
+ "iː˧": 199,
298
+ "ra˨˩.": 200,
299
+ "kʰa": 201,
300
+ "bon˧.": 202,
301
+ "ตำบล": 203,
302
+ "tam˧.bon˧.": 204,
303
+ "t˦˥.": 205,
304
+ "na˦˥.": 206,
305
+ "ɔːn˧": 207,
306
+ "at˦˥.": 208,
307
+ "n˩˩˦.": 209,
308
+ "sa": 210,
309
+ "ŋ˥˩.": 211,
310
+ "อน": 212,
311
+ "sɔː": 213,
312
+ "saː": 214,
313
+ "นา": 215,
314
+ "waː": 216,
315
+ "baːn˥˩.n": 217,
316
+ "ร์": 218,
317
+ "รร": 219,
318
+ "ɤː": 220,
319
+ "วั": 221,
320
+ "่ง": 222,
321
+ "t˨˩.": 223,
322
+ "าง": 224,
323
+ "tɕʰa˦˥.": 225,
324
+ "su": 226,
325
+ "ัก": 227,
326
+ "t˦˥": 228,
327
+ "k˦˥.": 229,
328
+ "ริ": 230,
329
+ "eʔ": 231,
330
+ "าย": 232,
331
+ "ติ": 233,
332
+ "คว": 234,
333
+ "aːj": 235,
334
+ "aj˧.": 236,
335
+ "an˧": 237,
336
+ "k˨˩": 238,
337
+ "สุ": 239,
338
+ "ɔːŋ˩˩˦.": 240,
339
+ "วิ": 241,
340
+ "at˦˥": 242,
341
+ "naː": 243,
342
+ "m˩˩˦.": 244,
343
+ "p˨˩.": 245,
344
+ "oː˧.": 246,
345
+ "ประ": 247,
346
+ "pra˨˩.": 248,
347
+ "aːn˧": 249,
348
+ "tɕʰa": 250,
349
+ "าก": 251,
350
+ "n˩˩˦": 252,
351
+ "หม": 253,
352
+ "k˥˩.": 254,
353
+ "eː˧.": 255,
354
+ "kʰwaː": 256,
355
+ "ภา": 257,
356
+ "͡ɕ": 258,
357
+ "ือ": 259,
358
+ "t͡ɕ": 260,
359
+ "ma˦˥.": 261,
360
+ "ี่": 262,
361
+ "aj˧": 263,
362
+ "เก": 264,
363
+ "tʰaː": 265,
364
+ "น์": 266,
365
+ "หล": 267,
366
+ "lɔː": 268,
367
+ "ɯː": 269,
368
+ "้น": 270,
369
+ "aj˥˩.": 271,
370
+ "ัง": 272,
371
+ "no": 273,
372
+ "ความ": 274,
373
+ "พิ": 275,
374
+ "ta˨˩.": 276,
375
+ "ซอย": 277,
376
+ "la": 278,
377
+ "ณ์": 279,
378
+ "̚˨˩.": 280,
379
+ "sɔːj˧.": 281,
380
+ "ภั": 282,
381
+ "หนอง": 283,
382
+ "t˨˩": 284,
383
+ "kʰwaːm˧.": 285,
384
+ "ิต": 286,
385
+ "at˨˩.": 287,
386
+ "pʰo": 288,
387
+ "ŋ˩˩˦": 289,
388
+ "ɔː˧.": 290,
389
+ "it˦˥.": 291,
390
+ "wa˦˥.": 292,
391
+ "พร": 293,
392
+ "น้": 294,
393
+ "พั": 295,
394
+ "ย์": 296,
395
+ "คล": 297,
396
+ "เส": 298,
397
+ "ื่": 299,
398
+ "ra": 300,
399
+ "jaː˧.": 301,
400
+ "ดา": 302,
401
+ "อก": 303,
402
+ "in˧.": 304,
403
+ "pʰuː": 305,
404
+ "ธิ": 306,
405
+ "pʰat˦˥.": 307,
406
+ "คร": 308,
407
+ "ɔːn˧.": 309,
408
+ "เข": 310,
409
+ "baːn˥˩.kʰ": 311,
410
+ "้ว": 312,
411
+ "ชา": 313,
412
+ "na": 314,
413
+ "pʰa˦˥.": 315,
414
+ "ัว": 316,
415
+ "su˨˩.": 317,
416
+ "so": 318,
417
+ "เล": 319,
418
+ "kʰaː": 320,
419
+ "kʰr": 321,
420
+ "ยา": 322,
421
+ "นิ": 323,
422
+ "tʰa˨˩.": 324,
423
+ "ม่": 325,
424
+ "ัญ": 326,
425
+ "กร": 327,
426
+ "ja˦˥.": 328,
427
+ "เท": 329,
428
+ "j˧": 330,
429
+ "ni˦˥.": 331,
430
+ "k˦˥": 332,
431
+ "วง": 333,
432
+ "บ้านหนอง": 334,
433
+ "naː˧.": 335,
434
+ "ัย": 336,
435
+ "baːn˥˩.nɔːŋ˩˩˦.": 337,
436
+ "pa˨˩.": 338,
437
+ "เป": 339,
438
+ "it˨˩.": 340,
439
+ "ŋ˨˩.": 341,
440
+ "รี": 342,
441
+ "siː": 343,
442
+ "ŋ˥˩": 344,
443
+ "jaː˧": 345,
444
+ "ka˨˩.": 346,
445
+ "n˥˩.": 347,
446
+ "pʰat˦˥": 348,
447
+ "ลา": 349,
448
+ "it˦˥": 350,
449
+ "p˦˥.": 351,
450
+ "si˨˩.": 352,
451
+ "สิ": 353,
452
+ "้อ": 354,
453
+ "ญา": 355,
454
+ "in˧": 356,
455
+ "kr": 357,
456
+ "wi˦˥.": 358,
457
+ "ณั": 359,
458
+ "oŋ˧.": 360,
459
+ "saː˩˩˦.": 361,
460
+ "m˦˥.": 362,
461
+ "tʰu": 363,
462
+ "aːŋ˧.": 364,
463
+ "เม": 365,
464
+ "บ้านโ": 366,
465
+ "pʰɔːn˧": 367,
466
+ "เร": 368,
467
+ "ับ": 369,
468
+ "้ง": 370,
469
+ "tʰi˦˥.": 371,
470
+ "aw": 372,
471
+ "เจ": 373,
472
+ "nat˦˥.": 374,
473
+ "la˦˥.": 375,
474
+ "on˧": 376,
475
+ "daː˧": 377,
476
+ "k̚": 378,
477
+ "่น": 379,
478
+ "pʰaː˧.": 380,
479
+ "ʔa˨˩.": 381,
480
+ "ti˨˩.": 382,
481
+ "ตา": 383,
482
+ "iː˥˩.": 384,
483
+ "กล": 385,
484
+ "สม": 386,
485
+ "raː˧.": 387,
486
+ "ตร": 388,
487
+ "ันท": 389,
488
+ "เด": 390,
489
+ "tʰɔː": 391,
490
+ "ru": 392,
491
+ "t˥˩": 393,
492
+ "น้ำ": 394,
493
+ "ทร": 395,
494
+ "ู้": 396,
495
+ "pʰi˦˥.": 397,
496
+ "วน": 398,
497
+ "ต์": 399,
498
+ "u˨˩.": 400,
499
+ "aŋ˧.": 401,
500
+ "ัด": 402,
501
+ "aːt˥˩": 403,
502
+ "ชน": 404,
503
+ "si": 405,
504
+ "it˨˩": 406,
505
+ "uː˧.": 407,
506
+ "บ้านท": 408,
507
+ "พล": 409,
508
+ "kra˨˩.": 410,
509
+ "ro": 411,
510
+ "กา": 412,
511
+ "t˥˩.": 413,
512
+ "laː": 414,
513
+ "นน": 415,
514
+ "maj": 416,
515
+ "rɔː": 417,
516
+ "naː˧": 418,
517
+ "paː": 419,
518
+ "roː": 420,
519
+ "มา": 421,
520
+ "aːt˥˩.": 422,
521
+ "pʰaː": 423,
522
+ "่อ": 424,
523
+ "โร": 425,
524
+ "ุม": 426,
525
+ "กระ": 427,
526
+ "p˥˩.": 428,
527
+ "ʔaː˧.": 429,
528
+ "baːn˥˩.h": 430,
529
+ "n˥˩": 431,
530
+ "ɔʔ": 432,
531
+ "t̚": 433,
532
+ "tɕa˨˩.": 434,
533
+ "kan˧.": 435,
534
+ "บ้านห": 436,
535
+ "็น": 437,
536
+ "nu": 438,
537
+ "บ้านส": 439,
538
+ "mɛː": 440,
539
+ "ที่": 441,
540
+ "ri˦˥.": 442,
541
+ "nɔː": 443,
542
+ "k˥˩": 444,
543
+ "seː": 445,
544
+ "un˧.": 446,
545
+ "รรณ": 447,
546
+ "p˦˥": 448,
547
+ "rat˦˥": 449,
548
+ "w˧.": 450,
549
+ "เน": 451,
550
+ "บุ": 452,
551
+ "kʰa˦˥.": 453,
552
+ "p˨˩": 454,
553
+ "lɔːŋ˧.": 455,
554
+ "kʰɔː": 456,
555
+ "ศิ": 457,
556
+ "ต้": 458,
557
+ "คลอง": 459,
558
+ "kaːn˧": 460,
559
+ "baːn˥˩.b": 461,
560
+ "rat˦˥.": 462,
561
+ "ดี": 463,
562
+ "pr": 464,
563
+ "siː˩˩˦.": 465,
564
+ "บ้านแ": 466,
565
+ "ินท": 467,
566
+ "พัฒ": 468,
567
+ "เช": 469,
568
+ "kʰaw": 470,
569
+ "นก": 471,
570
+ "kɛː": 472,
571
+ "tʰiː˥˩.": 473,
572
+ "เพ": 474,
573
+ "้วย": 475,
574
+ "m˥˩.": 476,
575
+ "ี้": 477,
576
+ "รรม": 478,
577
+ "baːn˥˩.d": 479,
578
+ "tʰi": 480,
579
+ "taː˧.": 481,
580
+ "aːt˨˩.": 482,
581
+ "n˨˩.": 483,
582
+ "kɔːn˧": 484,
583
+ "pʰuː˥˩.": 485,
584
+ "ิด": 486,
585
+ "บ้านป": 487,
586
+ "่าง": 488,
587
+ "aj˥˩": 489,
588
+ "รณ์": 490,
589
+ "aːt˨˩": 491,
590
+ "คำ": 492,
591
+ "ตะ": 493,
592
+ "อม": 494,
593
+ "ดง": 495,
594
+ "ุ่ง": 496,
595
+ "้าง": 497,
596
+ "เบ": 498,
597
+ "jaː": 499
598
+ },
599
+ "merges": [
600
+ [
601
+ "˩",
602
+ "."
603
+ ],
604
+ [
605
+ "˧",
606
+ "."
607
+ ],
608
+ [
609
+ "a",
610
+ "ː"
611
+ ],
612
+ [
613
+ "˦",
614
+ "˥"
615
+ ],
616
+ [
617
+ "˦˥",
618
+ "."
619
+ ],
620
+ [
621
+ "˨",
622
+ "˩."
623
+ ],
624
+ [
625
+ "˥",
626
+ "˩."
627
+ ],
628
+ [
629
+ "aː",
630
+ "n"
631
+ ],
632
+ [
633
+ "˩",
634
+ "˦"
635
+ ],
636
+ [
637
+ "˩",
638
+ "˩˦"
639
+ ],
640
+ [
641
+ "ʰ",
642
+ "a"
643
+ ],
644
+ [
645
+ "ɔ",
646
+ "ː"
647
+ ],
648
+ [
649
+ "้",
650
+ "า"
651
+ ],
652
+ [
653
+ "n",
654
+ "˧."
655
+ ],
656
+ [
657
+ "t",
658
+ "ɕ"
659
+ ],
660
+ [
661
+ "้า",
662
+ "น"
663
+ ],
664
+ [
665
+ "b",
666
+ "aːn"
667
+ ],
668
+ [
669
+ "บ",
670
+ "้าน"
671
+ ],
672
+ [
673
+ "baːn",
674
+ "˥˩."
675
+ ],
676
+ [
677
+ "˩˩˦",
678
+ "."
679
+ ],
680
+ [
681
+ "p",
682
+ "ʰ"
683
+ ],
684
+ [
685
+ "a",
686
+ "˨˩."
687
+ ],
688
+ [
689
+ "k",
690
+ "ʰ"
691
+ ],
692
+ [
693
+ "a",
694
+ "˦˥."
695
+ ],
696
+ [
697
+ "˨",
698
+ "˩"
699
+ ],
700
+ [
701
+ "t",
702
+ "ʰ"
703
+ ],
704
+ [
705
+ "aː",
706
+ "˧."
707
+ ],
708
+ [
709
+ "n",
710
+ "˧"
711
+ ],
712
+ [
713
+ "m",
714
+ "˧."
715
+ ],
716
+ [
717
+ "i",
718
+ "ː"
719
+ ],
720
+ [
721
+ "˥",
722
+ "˩"
723
+ ],
724
+ [
725
+ "ŋ",
726
+ "˧."
727
+ ],
728
+ [
729
+ "a",
730
+ "j"
731
+ ],
732
+ [
733
+ "ʰa",
734
+ "˦˥."
735
+ ],
736
+ [
737
+ "a",
738
+ "t"
739
+ ],
740
+ [
741
+ "ɛ",
742
+ "ː"
743
+ ],
744
+ [
745
+ "า",
746
+ "ร"
747
+ ],
748
+ [
749
+ "e",
750
+ "ː"
751
+ ],
752
+ [
753
+ "อ",
754
+ "ง"
755
+ ],
756
+ [
757
+ "aː",
758
+ "˧"
759
+ ],
760
+ [
761
+ "o",
762
+ "ː"
763
+ ],
764
+ [
765
+ "i",
766
+ "˦˥."
767
+ ],
768
+ [
769
+ "tɕ",
770
+ "ʰ"
771
+ ],
772
+ [
773
+ "u",
774
+ "ː"
775
+ ],
776
+ [
777
+ "i",
778
+ "t"
779
+ ],
780
+ [
781
+ "aːn",
782
+ "˧."
783
+ ],
784
+ [
785
+ "s",
786
+ "a˨˩."
787
+ ],
788
+ [
789
+ "p",
790
+ "ʰa"
791
+ ],
792
+ [
793
+ "a",
794
+ "m˧."
795
+ ],
796
+ [
797
+ "ŋ",
798
+ "˧"
799
+ ],
800
+ [
801
+ "o",
802
+ "n˧."
803
+ ],
804
+ [
805
+ "ร",
806
+ "ะ"
807
+ ],
808
+ [
809
+ "ั",
810
+ "น"
811
+ ],
812
+ [
813
+ "ก",
814
+ "าร"
815
+ ],
816
+ [
817
+ "u",
818
+ "a"
819
+ ],
820
+ [
821
+ "t",
822
+ "ʰa"
823
+ ],
824
+ [
825
+ "่",
826
+ "า"
827
+ ],
828
+ [
829
+ "r",
830
+ "a˦˥."
831
+ ],
832
+ [
833
+ "k",
834
+ "aːn˧."
835
+ ],
836
+ [
837
+ "a",
838
+ "n˧."
839
+ ],
840
+ [
841
+ "ิ",
842
+ "น"
843
+ ],
844
+ [
845
+ "i",
846
+ "˨˩."
847
+ ],
848
+ [
849
+ "ŋ",
850
+ "˩˩˦."
851
+ ],
852
+ [
853
+ "j",
854
+ "˧."
855
+ ],
856
+ [
857
+ "aː",
858
+ "t"
859
+ ],
860
+ [
861
+ "ห",
862
+ "น"
863
+ ],
864
+ [
865
+ "า",
866
+ "ม"
867
+ ],
868
+ [
869
+ "อ",
870
+ "ย"
871
+ ],
872
+ [
873
+ "ี",
874
+ "ย"
875
+ ],
876
+ [
877
+ "iː",
878
+ "˧."
879
+ ],
880
+ [
881
+ "i",
882
+ "a"
883
+ ],
884
+ [
885
+ "ɯ",
886
+ "a"
887
+ ],
888
+ [
889
+ "ร",
890
+ "า"
891
+ ],
892
+ [
893
+ "ร",
894
+ "ั"
895
+ ],
896
+ [
897
+ "m",
898
+ "˧"
899
+ ],
900
+ [
901
+ "k",
902
+ "˨˩."
903
+ ],
904
+ [
905
+ "ต",
906
+ "ำ"
907
+ ],
908
+ [
909
+ "t",
910
+ "am˧."
911
+ ],
912
+ [
913
+ "บ",
914
+ "ล"
915
+ ],
916
+ [
917
+ "า",
918
+ "น"
919
+ ],
920
+ [
921
+ "t",
922
+ "ʰa˦˥."
923
+ ],
924
+ [
925
+ "iː",
926
+ "˧"
927
+ ],
928
+ [
929
+ "r",
930
+ "a˨˩."
931
+ ],
932
+ [
933
+ "k",
934
+ "ʰa"
935
+ ],
936
+ [
937
+ "b",
938
+ "on˧."
939
+ ],
940
+ [
941
+ "ตำ",
942
+ "บล"
943
+ ],
944
+ [
945
+ "tam˧.",
946
+ "bon˧."
947
+ ],
948
+ [
949
+ "t",
950
+ "˦˥."
951
+ ],
952
+ [
953
+ "n",
954
+ "a˦˥."
955
+ ],
956
+ [
957
+ "ɔː",
958
+ "n˧"
959
+ ],
960
+ [
961
+ "at",
962
+ "˦˥."
963
+ ],
964
+ [
965
+ "n",
966
+ "˩˩˦."
967
+ ],
968
+ [
969
+ "s",
970
+ "a"
971
+ ],
972
+ [
973
+ "ŋ",
974
+ "˥˩."
975
+ ],
976
+ [
977
+ "อ",
978
+ "น"
979
+ ],
980
+ [
981
+ "s",
982
+ "ɔː"
983
+ ],
984
+ [
985
+ "s",
986
+ "aː"
987
+ ],
988
+ [
989
+ "น",
990
+ "า"
991
+ ],
992
+ [
993
+ "w",
994
+ "aː"
995
+ ],
996
+ [
997
+ "baːn˥˩.",
998
+ "n"
999
+ ],
1000
+ [
1001
+ "ร",
1002
+ "์"
1003
+ ],
1004
+ [
1005
+ "ร",
1006
+ "ร"
1007
+ ],
1008
+ [
1009
+ "ɤ",
1010
+ "ː"
1011
+ ],
1012
+ [
1013
+ "ว",
1014
+ "ั"
1015
+ ],
1016
+ [
1017
+ "่",
1018
+ "ง"
1019
+ ],
1020
+ [
1021
+ "t",
1022
+ "˨˩."
1023
+ ],
1024
+ [
1025
+ "า",
1026
+ "ง"
1027
+ ],
1028
+ [
1029
+ "tɕ",
1030
+ "ʰa˦˥."
1031
+ ],
1032
+ [
1033
+ "s",
1034
+ "u"
1035
+ ],
1036
+ [
1037
+ "ั",
1038
+ "ก"
1039
+ ],
1040
+ [
1041
+ "t",
1042
+ "˦˥"
1043
+ ],
1044
+ [
1045
+ "k",
1046
+ "˦˥."
1047
+ ],
1048
+ [
1049
+ "ร",
1050
+ "ิ"
1051
+ ],
1052
+ [
1053
+ "e",
1054
+ "ʔ"
1055
+ ],
1056
+ [
1057
+ "า",
1058
+ "ย"
1059
+ ],
1060
+ [
1061
+ "ต",
1062
+ "ิ"
1063
+ ],
1064
+ [
1065
+ "ค",
1066
+ "ว"
1067
+ ],
1068
+ [
1069
+ "aː",
1070
+ "j"
1071
+ ],
1072
+ [
1073
+ "aj",
1074
+ "˧."
1075
+ ],
1076
+ [
1077
+ "a",
1078
+ "n˧"
1079
+ ],
1080
+ [
1081
+ "k",
1082
+ "˨˩"
1083
+ ],
1084
+ [
1085
+ "ส",
1086
+ "ุ"
1087
+ ],
1088
+ [
1089
+ "ɔː",
1090
+ "ŋ˩˩˦."
1091
+ ],
1092
+ [
1093
+ "ว",
1094
+ "ิ"
1095
+ ],
1096
+ [
1097
+ "at",
1098
+ "˦˥"
1099
+ ],
1100
+ [
1101
+ "n",
1102
+ "aː"
1103
+ ],
1104
+ [
1105
+ "m",
1106
+ "˩˩˦."
1107
+ ],
1108
+ [
1109
+ "p",
1110
+ "˨˩."
1111
+ ],
1112
+ [
1113
+ "oː",
1114
+ "˧."
1115
+ ],
1116
+ [
1117
+ "ป",
1118
+ "ระ"
1119
+ ],
1120
+ [
1121
+ "p",
1122
+ "ra˨˩."
1123
+ ],
1124
+ [
1125
+ "aːn",
1126
+ "˧"
1127
+ ],
1128
+ [
1129
+ "tɕ",
1130
+ "ʰa"
1131
+ ],
1132
+ [
1133
+ "า",
1134
+ "ก"
1135
+ ],
1136
+ [
1137
+ "n",
1138
+ "˩˩˦"
1139
+ ],
1140
+ [
1141
+ "ห",
1142
+ "ม"
1143
+ ],
1144
+ [
1145
+ "k",
1146
+ "˥˩."
1147
+ ],
1148
+ [
1149
+ "eː",
1150
+ "˧."
1151
+ ],
1152
+ [
1153
+ "kʰ",
1154
+ "waː"
1155
+ ],
1156
+ [
1157
+ "ภ",
1158
+ "า"
1159
+ ],
1160
+ [
1161
+ "͡",
1162
+ "ɕ"
1163
+ ],
1164
+ [
1165
+ "ื",
1166
+ "อ"
1167
+ ],
1168
+ [
1169
+ "t",
1170
+ "͡ɕ"
1171
+ ],
1172
+ [
1173
+ "m",
1174
+ "a˦˥."
1175
+ ],
1176
+ [
1177
+ "ี",
1178
+ "่"
1179
+ ],
1180
+ [
1181
+ "aj",
1182
+ "˧"
1183
+ ],
1184
+ [
1185
+ "เ",
1186
+ "ก"
1187
+ ],
1188
+ [
1189
+ "tʰ",
1190
+ "aː"
1191
+ ],
1192
+ [
1193
+ "น",
1194
+ "์"
1195
+ ],
1196
+ [
1197
+ "ห",
1198
+ "ล"
1199
+ ],
1200
+ [
1201
+ "l",
1202
+ "ɔː"
1203
+ ],
1204
+ [
1205
+ "ɯ",
1206
+ "ː"
1207
+ ],
1208
+ [
1209
+ "้",
1210
+ "น"
1211
+ ],
1212
+ [
1213
+ "aj",
1214
+ "˥˩."
1215
+ ],
1216
+ [
1217
+ "ั",
1218
+ "ง"
1219
+ ],
1220
+ [
1221
+ "n",
1222
+ "o"
1223
+ ],
1224
+ [
1225
+ "คว",
1226
+ "าม"
1227
+ ],
1228
+ [
1229
+ "พ",
1230
+ "ิ"
1231
+ ],
1232
+ [
1233
+ "t",
1234
+ "a˨˩."
1235
+ ],
1236
+ [
1237
+ "ซ",
1238
+ "อย"
1239
+ ],
1240
+ [
1241
+ "l",
1242
+ "a"
1243
+ ],
1244
+ [
1245
+ "ณ",
1246
+ "์"
1247
+ ],
1248
+ [
1249
+ "̚",
1250
+ "˨˩."
1251
+ ],
1252
+ [
1253
+ "sɔː",
1254
+ "j˧."
1255
+ ],
1256
+ [
1257
+ "ภ",
1258
+ "ั"
1259
+ ],
1260
+ [
1261
+ "หน",
1262
+ "อง"
1263
+ ],
1264
+ [
1265
+ "t",
1266
+ "˨˩"
1267
+ ],
1268
+ [
1269
+ "kʰwaː",
1270
+ "m˧."
1271
+ ],
1272
+ [
1273
+ "ิ",
1274
+ "ต"
1275
+ ],
1276
+ [
1277
+ "at",
1278
+ "˨˩."
1279
+ ],
1280
+ [
1281
+ "pʰ",
1282
+ "o"
1283
+ ],
1284
+ [
1285
+ "ŋ",
1286
+ "˩˩˦"
1287
+ ],
1288
+ [
1289
+ "ɔː",
1290
+ "˧."
1291
+ ],
1292
+ [
1293
+ "it",
1294
+ "˦˥."
1295
+ ],
1296
+ [
1297
+ "w",
1298
+ "a˦˥."
1299
+ ],
1300
+ [
1301
+ "พ",
1302
+ "ร"
1303
+ ],
1304
+ [
1305
+ "น",
1306
+ "้"
1307
+ ],
1308
+ [
1309
+ "พ",
1310
+ "ั"
1311
+ ],
1312
+ [
1313
+ "ย",
1314
+ "์"
1315
+ ],
1316
+ [
1317
+ "ค",
1318
+ "ล"
1319
+ ],
1320
+ [
1321
+ "เ",
1322
+ "ส"
1323
+ ],
1324
+ [
1325
+ "ื",
1326
+ "่"
1327
+ ],
1328
+ [
1329
+ "r",
1330
+ "a"
1331
+ ],
1332
+ [
1333
+ "j",
1334
+ "aː˧."
1335
+ ],
1336
+ [
1337
+ "ด",
1338
+ "า"
1339
+ ],
1340
+ [
1341
+ "อ",
1342
+ "ก"
1343
+ ],
1344
+ [
1345
+ "i",
1346
+ "n˧."
1347
+ ],
1348
+ [
1349
+ "pʰ",
1350
+ "uː"
1351
+ ],
1352
+ [
1353
+ "ธ",
1354
+ "ิ"
1355
+ ],
1356
+ [
1357
+ "pʰa",
1358
+ "t˦˥."
1359
+ ],
1360
+ [
1361
+ "ค",
1362
+ "ร"
1363
+ ],
1364
+ [
1365
+ "ɔː",
1366
+ "n˧."
1367
+ ],
1368
+ [
1369
+ "เ",
1370
+ "ข"
1371
+ ],
1372
+ [
1373
+ "baːn˥˩.",
1374
+ "kʰ"
1375
+ ],
1376
+ [
1377
+ "้",
1378
+ "ว"
1379
+ ],
1380
+ [
1381
+ "ช",
1382
+ "า"
1383
+ ],
1384
+ [
1385
+ "n",
1386
+ "a"
1387
+ ],
1388
+ [
1389
+ "p",
1390
+ "ʰa˦˥."
1391
+ ],
1392
+ [
1393
+ "ั",
1394
+ "ว"
1395
+ ],
1396
+ [
1397
+ "su",
1398
+ "˨˩."
1399
+ ],
1400
+ [
1401
+ "s",
1402
+ "o"
1403
+ ],
1404
+ [
1405
+ "เ",
1406
+ "ล"
1407
+ ],
1408
+ [
1409
+ "kʰ",
1410
+ "aː"
1411
+ ],
1412
+ [
1413
+ "kʰ",
1414
+ "r"
1415
+ ],
1416
+ [
1417
+ "ย",
1418
+ "า"
1419
+ ],
1420
+ [
1421
+ "น",
1422
+ "ิ"
1423
+ ],
1424
+ [
1425
+ "tʰa",
1426
+ "˨˩."
1427
+ ],
1428
+ [
1429
+ "ม",
1430
+ "่"
1431
+ ],
1432
+ [
1433
+ "ั",
1434
+ "ญ"
1435
+ ],
1436
+ [
1437
+ "ก",
1438
+ "ร"
1439
+ ],
1440
+ [
1441
+ "j",
1442
+ "a˦˥."
1443
+ ],
1444
+ [
1445
+ "เ",
1446
+ "ท"
1447
+ ],
1448
+ [
1449
+ "j",
1450
+ "˧"
1451
+ ],
1452
+ [
1453
+ "n",
1454
+ "i˦˥."
1455
+ ],
1456
+ [
1457
+ "k",
1458
+ "˦˥"
1459
+ ],
1460
+ [
1461
+ "ว",
1462
+ "ง"
1463
+ ],
1464
+ [
1465
+ "บ้าน",
1466
+ "หนอง"
1467
+ ],
1468
+ [
1469
+ "n",
1470
+ "aː˧."
1471
+ ],
1472
+ [
1473
+ "ั",
1474
+ "ย"
1475
+ ],
1476
+ [
1477
+ "baːn˥˩.n",
1478
+ "ɔːŋ˩˩˦."
1479
+ ],
1480
+ [
1481
+ "p",
1482
+ "a˨˩."
1483
+ ],
1484
+ [
1485
+ "เ",
1486
+ "ป"
1487
+ ],
1488
+ [
1489
+ "it",
1490
+ "˨˩."
1491
+ ],
1492
+ [
1493
+ "ŋ",
1494
+ "˨˩."
1495
+ ],
1496
+ [
1497
+ "ร",
1498
+ "ี"
1499
+ ],
1500
+ [
1501
+ "s",
1502
+ "iː"
1503
+ ],
1504
+ [
1505
+ "ŋ",
1506
+ "˥˩"
1507
+ ],
1508
+ [
1509
+ "j",
1510
+ "aː˧"
1511
+ ],
1512
+ [
1513
+ "k",
1514
+ "a˨˩."
1515
+ ],
1516
+ [
1517
+ "n",
1518
+ "˥˩."
1519
+ ],
1520
+ [
1521
+ "pʰa",
1522
+ "t˦˥"
1523
+ ],
1524
+ [
1525
+ "ล",
1526
+ "า"
1527
+ ],
1528
+ [
1529
+ "it",
1530
+ "˦˥"
1531
+ ],
1532
+ [
1533
+ "p",
1534
+ "˦˥."
1535
+ ],
1536
+ [
1537
+ "s",
1538
+ "i˨˩."
1539
+ ],
1540
+ [
1541
+ "ส",
1542
+ "ิ"
1543
+ ],
1544
+ [
1545
+ "้",
1546
+ "อ"
1547
+ ],
1548
+ [
1549
+ "ญ",
1550
+ "า"
1551
+ ],
1552
+ [
1553
+ "i",
1554
+ "n˧"
1555
+ ],
1556
+ [
1557
+ "k",
1558
+ "r"
1559
+ ],
1560
+ [
1561
+ "w",
1562
+ "i˦˥."
1563
+ ],
1564
+ [
1565
+ "ณ",
1566
+ "ั"
1567
+ ],
1568
+ [
1569
+ "o",
1570
+ "ŋ˧."
1571
+ ],
1572
+ [
1573
+ "saː",
1574
+ "˩˩˦."
1575
+ ],
1576
+ [
1577
+ "m",
1578
+ "˦˥."
1579
+ ],
1580
+ [
1581
+ "tʰ",
1582
+ "u"
1583
+ ],
1584
+ [
1585
+ "aː",
1586
+ "ŋ˧."
1587
+ ],
1588
+ [
1589
+ "เ",
1590
+ "ม"
1591
+ ],
1592
+ [
1593
+ "บ้าน",
1594
+ "โ"
1595
+ ],
1596
+ [
1597
+ "pʰ",
1598
+ "ɔːn˧"
1599
+ ],
1600
+ [
1601
+ "เ",
1602
+ "ร"
1603
+ ],
1604
+ [
1605
+ "ั",
1606
+ "บ"
1607
+ ],
1608
+ [
1609
+ "้",
1610
+ "ง"
1611
+ ],
1612
+ [
1613
+ "tʰ",
1614
+ "i˦˥."
1615
+ ],
1616
+ [
1617
+ "a",
1618
+ "w"
1619
+ ],
1620
+ [
1621
+ "เ",
1622
+ "จ"
1623
+ ],
1624
+ [
1625
+ "n",
1626
+ "at˦˥."
1627
+ ],
1628
+ [
1629
+ "l",
1630
+ "a˦˥."
1631
+ ],
1632
+ [
1633
+ "o",
1634
+ "n˧"
1635
+ ],
1636
+ [
1637
+ "d",
1638
+ "aː˧"
1639
+ ],
1640
+ [
1641
+ "k",
1642
+ "̚"
1643
+ ],
1644
+ [
1645
+ "่",
1646
+ "น"
1647
+ ],
1648
+ [
1649
+ "pʰ",
1650
+ "aː˧."
1651
+ ],
1652
+ [
1653
+ "ʔ",
1654
+ "a˨˩."
1655
+ ],
1656
+ [
1657
+ "t",
1658
+ "i˨˩."
1659
+ ],
1660
+ [
1661
+ "ต",
1662
+ "า"
1663
+ ],
1664
+ [
1665
+ "iː",
1666
+ "˥˩."
1667
+ ],
1668
+ [
1669
+ "ก",
1670
+ "ล"
1671
+ ],
1672
+ [
1673
+ "ส",
1674
+ "ม"
1675
+ ],
1676
+ [
1677
+ "r",
1678
+ "aː˧."
1679
+ ],
1680
+ [
1681
+ "ต",
1682
+ "ร"
1683
+ ],
1684
+ [
1685
+ "ัน",
1686
+ "ท"
1687
+ ],
1688
+ [
1689
+ "เ",
1690
+ "ด"
1691
+ ],
1692
+ [
1693
+ "tʰ",
1694
+ "ɔː"
1695
+ ],
1696
+ [
1697
+ "r",
1698
+ "u"
1699
+ ],
1700
+ [
1701
+ "t",
1702
+ "˥˩"
1703
+ ],
1704
+ [
1705
+ "น้",
1706
+ "ำ"
1707
+ ],
1708
+ [
1709
+ "ท",
1710
+ "ร"
1711
+ ],
1712
+ [
1713
+ "ู",
1714
+ "้"
1715
+ ],
1716
+ [
1717
+ "pʰ",
1718
+ "i˦˥."
1719
+ ],
1720
+ [
1721
+ "ว",
1722
+ "น"
1723
+ ],
1724
+ [
1725
+ "ต",
1726
+ "์"
1727
+ ],
1728
+ [
1729
+ "u",
1730
+ "˨˩."
1731
+ ],
1732
+ [
1733
+ "a",
1734
+ "ŋ˧."
1735
+ ],
1736
+ [
1737
+ "ั",
1738
+ "ด"
1739
+ ],
1740
+ [
1741
+ "aːt",
1742
+ "˥˩"
1743
+ ],
1744
+ [
1745
+ "ช",
1746
+ "น"
1747
+ ],
1748
+ [
1749
+ "s",
1750
+ "i"
1751
+ ],
1752
+ [
1753
+ "it",
1754
+ "˨˩"
1755
+ ],
1756
+ [
1757
+ "uː",
1758
+ "˧."
1759
+ ],
1760
+ [
1761
+ "บ้าน",
1762
+ "ท"
1763
+ ],
1764
+ [
1765
+ "พ",
1766
+ "ล"
1767
+ ],
1768
+ [
1769
+ "k",
1770
+ "ra˨˩."
1771
+ ],
1772
+ [
1773
+ "r",
1774
+ "o"
1775
+ ],
1776
+ [
1777
+ "ก",
1778
+ "า"
1779
+ ],
1780
+ [
1781
+ "t",
1782
+ "˥˩."
1783
+ ],
1784
+ [
1785
+ "l",
1786
+ "aː"
1787
+ ],
1788
+ [
1789
+ "น",
1790
+ "น"
1791
+ ],
1792
+ [
1793
+ "m",
1794
+ "aj"
1795
+ ],
1796
+ [
1797
+ "r",
1798
+ "ɔː"
1799
+ ],
1800
+ [
1801
+ "n",
1802
+ "aː˧"
1803
+ ],
1804
+ [
1805
+ "p",
1806
+ "aː"
1807
+ ],
1808
+ [
1809
+ "r",
1810
+ "oː"
1811
+ ],
1812
+ [
1813
+ "ม",
1814
+ "า"
1815
+ ],
1816
+ [
1817
+ "aːt",
1818
+ "˥˩."
1819
+ ],
1820
+ [
1821
+ "pʰ",
1822
+ "aː"
1823
+ ],
1824
+ [
1825
+ "่",
1826
+ "อ"
1827
+ ],
1828
+ [
1829
+ "โ",
1830
+ "ร"
1831
+ ],
1832
+ [
1833
+ "ุ",
1834
+ "ม"
1835
+ ],
1836
+ [
1837
+ "ก",
1838
+ "ระ"
1839
+ ],
1840
+ [
1841
+ "p",
1842
+ "˥˩."
1843
+ ],
1844
+ [
1845
+ "ʔ",
1846
+ "aː˧."
1847
+ ],
1848
+ [
1849
+ "baːn˥˩.",
1850
+ "h"
1851
+ ],
1852
+ [
1853
+ "n",
1854
+ "˥˩"
1855
+ ],
1856
+ [
1857
+ "ɔ",
1858
+ "ʔ"
1859
+ ],
1860
+ [
1861
+ "t",
1862
+ "̚"
1863
+ ],
1864
+ [
1865
+ "tɕ",
1866
+ "a˨˩."
1867
+ ],
1868
+ [
1869
+ "k",
1870
+ "an˧."
1871
+ ],
1872
+ [
1873
+ "บ้าน",
1874
+ "ห"
1875
+ ],
1876
+ [
1877
+ "็",
1878
+ "น"
1879
+ ],
1880
+ [
1881
+ "n",
1882
+ "u"
1883
+ ],
1884
+ [
1885
+ "บ้าน",
1886
+ "ส"
1887
+ ],
1888
+ [
1889
+ "m",
1890
+ "ɛː"
1891
+ ],
1892
+ [
1893
+ "ท",
1894
+ "ี่"
1895
+ ],
1896
+ [
1897
+ "r",
1898
+ "i˦˥."
1899
+ ],
1900
+ [
1901
+ "n",
1902
+ "ɔː"
1903
+ ],
1904
+ [
1905
+ "k",
1906
+ "˥˩"
1907
+ ],
1908
+ [
1909
+ "s",
1910
+ "eː"
1911
+ ],
1912
+ [
1913
+ "u",
1914
+ "n˧."
1915
+ ],
1916
+ [
1917
+ "รร",
1918
+ "ณ"
1919
+ ],
1920
+ [
1921
+ "p",
1922
+ "˦˥"
1923
+ ],
1924
+ [
1925
+ "r",
1926
+ "at˦˥"
1927
+ ],
1928
+ [
1929
+ "w",
1930
+ "˧."
1931
+ ],
1932
+ [
1933
+ "เ",
1934
+ "น"
1935
+ ],
1936
+ [
1937
+ "บ",
1938
+ "ุ"
1939
+ ],
1940
+ [
1941
+ "k",
1942
+ "ʰa˦˥."
1943
+ ],
1944
+ [
1945
+ "p",
1946
+ "˨˩"
1947
+ ],
1948
+ [
1949
+ "lɔː",
1950
+ "ŋ˧."
1951
+ ],
1952
+ [
1953
+ "kʰ",
1954
+ "ɔː"
1955
+ ],
1956
+ [
1957
+ "ศ",
1958
+ "ิ"
1959
+ ],
1960
+ [
1961
+ "ต",
1962
+ "้"
1963
+ ],
1964
+ [
1965
+ "คล",
1966
+ "อง"
1967
+ ],
1968
+ [
1969
+ "k",
1970
+ "aːn˧"
1971
+ ],
1972
+ [
1973
+ "baːn˥˩.",
1974
+ "b"
1975
+ ],
1976
+ [
1977
+ "r",
1978
+ "at˦˥."
1979
+ ],
1980
+ [
1981
+ "ด",
1982
+ "ี"
1983
+ ],
1984
+ [
1985
+ "p",
1986
+ "r"
1987
+ ],
1988
+ [
1989
+ "siː",
1990
+ "˩˩˦."
1991
+ ],
1992
+ [
1993
+ "บ้าน",
1994
+ "แ"
1995
+ ],
1996
+ [
1997
+ "ิน",
1998
+ "ท"
1999
+ ],
2000
+ [
2001
+ "พั",
2002
+ "ฒ"
2003
+ ],
2004
+ [
2005
+ "เ",
2006
+ "ช"
2007
+ ],
2008
+ [
2009
+ "kʰa",
2010
+ "w"
2011
+ ],
2012
+ [
2013
+ "น",
2014
+ "ก"
2015
+ ],
2016
+ [
2017
+ "k",
2018
+ "ɛː"
2019
+ ],
2020
+ [
2021
+ "tʰ",
2022
+ "iː˥˩."
2023
+ ],
2024
+ [
2025
+ "เ",
2026
+ "พ"
2027
+ ],
2028
+ [
2029
+ "้ว",
2030
+ "ย"
2031
+ ],
2032
+ [
2033
+ "m",
2034
+ "˥˩."
2035
+ ],
2036
+ [
2037
+ "ี",
2038
+ "้"
2039
+ ],
2040
+ [
2041
+ "รร",
2042
+ "ม"
2043
+ ],
2044
+ [
2045
+ "baːn˥˩.",
2046
+ "d"
2047
+ ],
2048
+ [
2049
+ "tʰ",
2050
+ "i"
2051
+ ],
2052
+ [
2053
+ "t",
2054
+ "aː˧."
2055
+ ],
2056
+ [
2057
+ "aːt",
2058
+ "˨˩."
2059
+ ],
2060
+ [
2061
+ "n",
2062
+ "˨˩."
2063
+ ],
2064
+ [
2065
+ "k",
2066
+ "ɔːn˧"
2067
+ ],
2068
+ [
2069
+ "pʰuː",
2070
+ "˥˩."
2071
+ ],
2072
+ [
2073
+ "ิ",
2074
+ "ด"
2075
+ ],
2076
+ [
2077
+ "บ้าน",
2078
+ "ป"
2079
+ ],
2080
+ [
2081
+ "่า",
2082
+ "ง"
2083
+ ],
2084
+ [
2085
+ "aj",
2086
+ "˥˩"
2087
+ ],
2088
+ [
2089
+ "ร",
2090
+ "ณ์"
2091
+ ],
2092
+ [
2093
+ "aːt",
2094
+ "˨˩"
2095
+ ],
2096
+ [
2097
+ "ค",
2098
+ "ำ"
2099
+ ],
2100
+ [
2101
+ "ต",
2102
+ "ะ"
2103
+ ],
2104
+ [
2105
+ "อ",
2106
+ "ม"
2107
+ ],
2108
+ [
2109
+ "ด",
2110
+ "ง"
2111
+ ],
2112
+ [
2113
+ "ุ",
2114
+ "่ง"
2115
+ ],
2116
+ [
2117
+ "้า",
2118
+ "ง"
2119
+ ],
2120
+ [
2121
+ "เ",
2122
+ "บ"
2123
+ ],
2124
+ [
2125
+ "j",
2126
+ "aː"
2127
+ ]
2128
+ ]
2129
+ }
2130
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<pad>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "</s>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "<s>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ }
35
+ },
36
+ "additional_special_tokens": [],
37
+ "bos_token": "<s>",
38
+ "clean_up_tokenization_spaces": false,
39
+ "eos_token": "</s>",
40
+ "extra_ids": 0,
41
+ "extra_special_tokens": {},
42
+ "model_max_length": 1000000000000000019884624838656,
43
+ "pad_token": "<pad>",
44
+ "tokenizer_class": "T5Tokenizer",
45
+ "unk_token": "<unk>"
46
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6229abc360e4edcf763f457e80798f82a2f28a57dd1acbbc42340a7e84774d29
3
+ size 7672