Hugging Face's logo

hf-internal-testing
/
tiny-processor-glmasr

glmasr
Eval Results
Model card Files
xet
 Community
hf-transformers-bot commited on
Commit
a26cd09
·
verified ·
1 Parent(s): 6ceee53

Add tiny processor fixture for testing

Browse files
Files changed (7) hide show
  1. .eval_results/open_asr_leaderboard.yaml +99 -0
  2. chat_template.jinja +32 -0
  3. config.json +60 -0
  4. generation_config.json +10 -0
  5. processor_config.json +20 -0
  6. tokenizer.json +1783 -0
  7. tokenizer_config.json +175 -0
.eval_results/open_asr_leaderboard.yaml ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ - dataset:
2
+ id: hf-audio/open-asr-leaderboard
3
+ task_id: mean_wer
4
+ value: 7.03
5
+ date: '2025-12-09'
6
+ source:
7
+ url: https://huggingface.co/hf-audio
8
+ name: open-asr-leaderboard
9
+ user: hf-audio
10
+
11
+ - dataset:
12
+ id: hf-audio/open-asr-leaderboard
13
+ task_id: rtfx
14
+ value: 145.28
15
+ date: '2025-12-09'
16
+ source:
17
+ url: https://huggingface.co/hf-audio
18
+ name: open-asr-leaderboard
19
+ user: hf-audio
20
+
21
+ - dataset:
22
+ id: hf-audio/open-asr-leaderboard
23
+ task_id: ami_wer
24
+ value: 16.15
25
+ date: '2025-12-09'
26
+ source:
27
+ url: https://huggingface.co/hf-audio
28
+ name: open-asr-leaderboard
29
+ user: hf-audio
30
+
31
+ - dataset:
32
+ id: hf-audio/open-asr-leaderboard
33
+ task_id: earnings22_wer
34
+ value: 11.08
35
+ date: '2025-12-09'
36
+ source:
37
+ url: https://huggingface.co/hf-audio
38
+ name: open-asr-leaderboard
39
+ user: hf-audio
40
+
41
+ - dataset:
42
+ id: hf-audio/open-asr-leaderboard
43
+ task_id: gigaspeech_wer
44
+ value: 9.73
45
+ date: '2025-12-09'
46
+ source:
47
+ url: https://huggingface.co/hf-audio
48
+ name: open-asr-leaderboard
49
+ user: hf-audio
50
+
51
+ - dataset:
52
+ id: hf-audio/open-asr-leaderboard
53
+ task_id: librispeech_clean_wer
54
+ value: 2.15
55
+ date: '2025-12-09'
56
+ source:
57
+ url: https://huggingface.co/hf-audio
58
+ name: open-asr-leaderboard
59
+ user: hf-audio
60
+
61
+ - dataset:
62
+ id: hf-audio/open-asr-leaderboard
63
+ task_id: librispeech_other_wer
64
+ value: 4.42
65
+ date: '2025-12-09'
66
+ source:
67
+ url: https://huggingface.co/hf-audio
68
+ name: open-asr-leaderboard
69
+ user: hf-audio
70
+
71
+ - dataset:
72
+ id: hf-audio/open-asr-leaderboard
73
+ task_id: spgispeech_wer
74
+ value: 2.08
75
+ date: '2025-12-09'
76
+ source:
77
+ url: https://huggingface.co/hf-audio
78
+ name: open-asr-leaderboard
79
+ user: hf-audio
80
+
81
+ - dataset:
82
+ id: hf-audio/open-asr-leaderboard
83
+ task_id: tedlium_wer
84
+ value: 3.1
85
+ date: '2025-12-09'
86
+ source:
87
+ url: https://huggingface.co/hf-audio
88
+ name: open-asr-leaderboard
89
+ user: hf-audio
90
+
91
+ - dataset:
92
+ id: hf-audio/open-asr-leaderboard
93
+ task_id: voxpopuli_wer
94
+ value: 7.54
95
+ date: '2025-12-09'
96
+ source:
97
+ url: https://huggingface.co/hf-audio
98
+ name: open-asr-leaderboard
99
+ user: hf-audio
chat_template.jinja ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- macro to_text(content) -%}
2
+ {%- if content is string -%}
3
+ {{- content -}}
4
+ {%- elif content is iterable and content is not mapping -%}
5
+ {%- for item in content -%}
6
+ {%- if item is mapping and item.type == 'text' and item.text is defined -%}
7
+ {{- item.text -}}
8
+ {%- elif item is mapping and (item.type == 'audio' or 'audio' in item) -%}
9
+ <|begin_of_audio|><|pad|><|end_of_audio|><|user|>
10
+ {% elif item is string -%}
11
+ {{- item -}}
12
+ {%- endif -%}
13
+ {%- endfor -%}
14
+ {%- else -%}
15
+ {{- content -}}
16
+ {%- endif -%}
17
+ {%- endmacro -%}
18
+ {%- for m in messages -%}
19
+ {%- if m.role == 'system' -%}
20
+ <|system|>
21
+ {{ to_text(m.content) | trim }}
22
+ {%- elif m.role == 'user' -%}
23
+ <|user|>
24
+ {{ to_text(m.content) | trim }}
25
+ {%- elif m.role == 'assistant' -%}
26
+ <|assistant|>
27
+ {{ to_text(m.content) | trim }}
28
+ {%- endif -%}
29
+ {%- endfor -%}
30
+ {%- if add_generation_prompt -%}
31
+ <|assistant|>
32
+ {% endif -%}
config.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "GlmAsrForConditionalGeneration"
4
+ ],
5
+ "audio_config": {
6
+ "attention_dropout": 0.0,
7
+ "head_dim": 64,
8
+ "hidden_act": "gelu",
9
+ "hidden_size": 1280,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 5120,
12
+ "max_position_embeddings": 1500,
13
+ "model_type": "glmasr_encoder",
14
+ "num_attention_heads": 20,
15
+ "num_hidden_layers": 32,
16
+ "num_key_value_heads": 20,
17
+ "num_mel_bins": 128,
18
+ "partial_rotary_factor": 0.5,
19
+ "rope_parameters": {
20
+ "partial_rotary_factor": 0.5,
21
+ "rope_theta": 10000.0,
22
+ "rope_type": "default"
23
+ }
24
+ },
25
+ "audio_token_id": 59260,
26
+ "dtype": "bfloat16",
27
+ "hidden_size": 2048,
28
+ "model_type": "glmasr",
29
+ "projector_hidden_act": "gelu",
30
+ "text_config": {
31
+ "attention_bias": false,
32
+ "attention_dropout": 0.0,
33
+ "eos_token_id": [
34
+ 59246,
35
+ 59253,
36
+ 59255
37
+ ],
38
+ "head_dim": 128,
39
+ "hidden_act": "silu",
40
+ "hidden_size": 2048,
41
+ "initializer_range": 0.02,
42
+ "intermediate_size": 6144,
43
+ "max_position_embeddings": 8192,
44
+ "mlp_bias": false,
45
+ "model_type": "llama",
46
+ "num_attention_heads": 16,
47
+ "num_hidden_layers": 28,
48
+ "num_key_value_heads": 4,
49
+ "pretraining_tp": 1,
50
+ "rms_norm_eps": 1e-05,
51
+ "rope_parameters": {
52
+ "rope_theta": 10000.0,
53
+ "rope_type": "default"
54
+ },
55
+ "use_cache": true,
56
+ "vocab_size": 59264
57
+ },
58
+ "transformers_version": "5.0.0.dev0",
59
+ "vocab_size": 500
60
+ }
generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": [
5
+ 59246,
6
+ 59253,
7
+ 59255
8
+ ],
9
+ "transformers_version": "5.0.0.dev0"
10
+ }
processor_config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "audio_token": "<|pad|>",
3
+ "default_transcription_prompt": "Please transcribe this audio into text",
4
+ "feature_extractor": {
5
+ "chunk_length": 30,
6
+ "dither": 0.0,
7
+ "feature_extractor_type": "WhisperFeatureExtractor",
8
+ "feature_size": 128,
9
+ "hop_length": 160,
10
+ "n_fft": 400,
11
+ "n_samples": 480000,
12
+ "nb_max_frames": 3000,
13
+ "padding_side": "right",
14
+ "padding_value": 0.0,
15
+ "return_attention_mask": false,
16
+ "sampling_rate": 16000
17
+ },
18
+ "max_audio_len": 655,
19
+ "processor_class": "GlmAsrProcessor"
20
+ }
tokenizer.json ADDED
@@ -0,0 +1,1783 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "1.0",
3
+ "truncation": null,
4
+ "padding": null,
5
+ "added_tokens": [
6
+ {
7
+ "id": 500,
8
+ "content": "<|endoftext|>",
9
+ "single_word": false,
10
+ "lstrip": false,
11
+ "rstrip": false,
12
+ "normalized": false,
13
+ "special": true
14
+ },
15
+ {
16
+ "id": 501,
17
+ "content": "[MASK]",
18
+ "single_word": false,
19
+ "lstrip": false,
20
+ "rstrip": false,
21
+ "normalized": false,
22
+ "special": true
23
+ },
24
+ {
25
+ "id": 502,
26
+ "content": "[gMASK]",
27
+ "single_word": false,
28
+ "lstrip": false,
29
+ "rstrip": false,
30
+ "normalized": false,
31
+ "special": true
32
+ },
33
+ {
34
+ "id": 503,
35
+ "content": "[sMASK]",
36
+ "single_word": false,
37
+ "lstrip": false,
38
+ "rstrip": false,
39
+ "normalized": false,
40
+ "special": true
41
+ },
42
+ {
43
+ "id": 504,
44
+ "content": "<sop>",
45
+ "single_word": false,
46
+ "lstrip": false,
47
+ "rstrip": false,
48
+ "normalized": false,
49
+ "special": true
50
+ },
51
+ {
52
+ "id": 505,
53
+ "content": "<eop>",
54
+ "single_word": false,
55
+ "lstrip": false,
56
+ "rstrip": false,
57
+ "normalized": false,
58
+ "special": true
59
+ },
60
+ {
61
+ "id": 506,
62
+ "content": "<|system|>",
63
+ "single_word": false,
64
+ "lstrip": false,
65
+ "rstrip": false,
66
+ "normalized": false,
67
+ "special": true
68
+ },
69
+ {
70
+ "id": 507,
71
+ "content": "<|user|>",
72
+ "single_word": false,
73
+ "lstrip": false,
74
+ "rstrip": false,
75
+ "normalized": false,
76
+ "special": true
77
+ },
78
+ {
79
+ "id": 508,
80
+ "content": "<|assistant|>",
81
+ "single_word": false,
82
+ "lstrip": false,
83
+ "rstrip": false,
84
+ "normalized": false,
85
+ "special": true
86
+ },
87
+ {
88
+ "id": 509,
89
+ "content": "<|observation|>",
90
+ "single_word": false,
91
+ "lstrip": false,
92
+ "rstrip": false,
93
+ "normalized": false,
94
+ "special": true
95
+ },
96
+ {
97
+ "id": 510,
98
+ "content": "<|begin_of_image|>",
99
+ "single_word": false,
100
+ "lstrip": false,
101
+ "rstrip": false,
102
+ "normalized": false,
103
+ "special": true
104
+ },
105
+ {
106
+ "id": 511,
107
+ "content": "<|end_of_image|>",
108
+ "single_word": false,
109
+ "lstrip": false,
110
+ "rstrip": false,
111
+ "normalized": false,
112
+ "special": true
113
+ },
114
+ {
115
+ "id": 512,
116
+ "content": "<|begin_of_video|>",
117
+ "single_word": false,
118
+ "lstrip": false,
119
+ "rstrip": false,
120
+ "normalized": false,
121
+ "special": true
122
+ },
123
+ {
124
+ "id": 513,
125
+ "content": "<|end_of_video|>",
126
+ "single_word": false,
127
+ "lstrip": false,
128
+ "rstrip": false,
129
+ "normalized": false,
130
+ "special": true
131
+ },
132
+ {
133
+ "id": 514,
134
+ "content": "<|pad|>",
135
+ "single_word": false,
136
+ "lstrip": false,
137
+ "rstrip": false,
138
+ "normalized": false,
139
+ "special": true
140
+ },
141
+ {
142
+ "id": 515,
143
+ "content": "<|begin_of_audio|>",
144
+ "single_word": false,
145
+ "lstrip": false,
146
+ "rstrip": false,
147
+ "normalized": false,
148
+ "special": true
149
+ },
150
+ {
151
+ "id": 516,
152
+ "content": "<|end_of_audio|>",
153
+ "single_word": false,
154
+ "lstrip": false,
155
+ "rstrip": false,
156
+ "normalized": false,
157
+ "special": true
158
+ }
159
+ ],
160
+ "normalizer": null,
161
+ "pre_tokenizer": {
162
+ "type": "Sequence",
163
+ "pretokenizers": [
164
+ {
165
+ "type": "Split",
166
+ "pattern": {
167
+ "Regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+"
168
+ },
169
+ "behavior": "Isolated",
170
+ "invert": false
171
+ },
172
+ {
173
+ "type": "ByteLevel",
174
+ "add_prefix_space": false,
175
+ "trim_offsets": true,
176
+ "use_regex": false
177
+ }
178
+ ]
179
+ },
180
+ "post_processor": {
181
+ "type": "Sequence",
182
+ "processors": [
183
+ {
184
+ "type": "ByteLevel",
185
+ "add_prefix_space": true,
186
+ "trim_offsets": false,
187
+ "use_regex": true
188
+ }
189
+ ]
190
+ },
191
+ "decoder": {
192
+ "type": "ByteLevel",
193
+ "add_prefix_space": true,
194
+ "trim_offsets": true,
195
+ "use_regex": true
196
+ },
197
+ "model": {
198
+ "type": "BPE",
199
+ "dropout": null,
200
+ "unk_token": null,
201
+ "continuing_subword_prefix": null,
202
+ "end_of_word_suffix": null,
203
+ "fuse_unk": false,
204
+ "byte_fallback": false,
205
+ "ignore_merges": true,
206
+ "vocab": {
207
+ "Ā": 0,
208
+ "ā": 1,
209
+ "Ă": 2,
210
+ "ă": 3,
211
+ "Ą": 4,
212
+ "ą": 5,
213
+ "Ć": 6,
214
+ "ć": 7,
215
+ "Ĉ": 8,
216
+ "ĉ": 9,
217
+ "Ċ": 10,
218
+ "ċ": 11,
219
+ "Č": 12,
220
+ "č": 13,
221
+ "Ď": 14,
222
+ "ď": 15,
223
+ "Đ": 16,
224
+ "đ": 17,
225
+ "Ē": 18,
226
+ "ē": 19,
227
+ "Ĕ": 20,
228
+ "ĕ": 21,
229
+ "Ė": 22,
230
+ "ė": 23,
231
+ "Ę": 24,
232
+ "ę": 25,
233
+ "Ě": 26,
234
+ "ě": 27,
235
+ "Ĝ": 28,
236
+ "ĝ": 29,
237
+ "Ğ": 30,
238
+ "ğ": 31,
239
+ "Ġ": 32,
240
+ "!": 33,
241
+ "\"": 34,
242
+ "#": 35,
243
+ "$": 36,
244
+ "%": 37,
245
+ "&": 38,
246
+ "'": 39,
247
+ "(": 40,
248
+ ")": 41,
249
+ "*": 42,
250
+ "+": 43,
251
+ ",": 44,
252
+ "-": 45,
253
+ ".": 46,
254
+ "/": 47,
255
+ "0": 48,
256
+ "1": 49,
257
+ "2": 50,
258
+ "3": 51,
259
+ "4": 52,
260
+ "5": 53,
261
+ "6": 54,
262
+ "7": 55,
263
+ "8": 56,
264
+ "9": 57,
265
+ ":": 58,
266
+ ";": 59,
267
+ "<": 60,
268
+ "=": 61,
269
+ ">": 62,
270
+ "?": 63,
271
+ "@": 64,
272
+ "A": 65,
273
+ "B": 66,
274
+ "C": 67,
275
+ "D": 68,
276
+ "E": 69,
277
+ "F": 70,
278
+ "G": 71,
279
+ "H": 72,
280
+ "I": 73,
281
+ "J": 74,
282
+ "K": 75,
283
+ "L": 76,
284
+ "M": 77,
285
+ "N": 78,
286
+ "O": 79,
287
+ "P": 80,
288
+ "Q": 81,
289
+ "R": 82,
290
+ "S": 83,
291
+ "T": 84,
292
+ "U": 85,
293
+ "V": 86,
294
+ "W": 87,
295
+ "X": 88,
296
+ "Y": 89,
297
+ "Z": 90,
298
+ "[": 91,
299
+ "\\": 92,
300
+ "]": 93,
301
+ "^": 94,
302
+ "_": 95,
303
+ "`": 96,
304
+ "a": 97,
305
+ "b": 98,
306
+ "c": 99,
307
+ "d": 100,
308
+ "e": 101,
309
+ "f": 102,
310
+ "g": 103,
311
+ "h": 104,
312
+ "i": 105,
313
+ "j": 106,
314
+ "k": 107,
315
+ "l": 108,
316
+ "m": 109,
317
+ "n": 110,
318
+ "o": 111,
319
+ "p": 112,
320
+ "q": 113,
321
+ "r": 114,
322
+ "s": 115,
323
+ "t": 116,
324
+ "u": 117,
325
+ "v": 118,
326
+ "w": 119,
327
+ "x": 120,
328
+ "y": 121,
329
+ "z": 122,
330
+ "{": 123,
331
+ "|": 124,
332
+ "}": 125,
333
+ "~": 126,
334
+ "ġ": 127,
335
+ "Ģ": 128,
336
+ "ģ": 129,
337
+ "Ĥ": 130,
338
+ "ĥ": 131,
339
+ "Ħ": 132,
340
+ "ħ": 133,
341
+ "Ĩ": 134,
342
+ "ĩ": 135,
343
+ "Ī": 136,
344
+ "ī": 137,
345
+ "Ĭ": 138,
346
+ "ĭ": 139,
347
+ "Į": 140,
348
+ "į": 141,
349
+ "İ": 142,
350
+ "ı": 143,
351
+ "IJ": 144,
352
+ "ij": 145,
353
+ "Ĵ": 146,
354
+ "ĵ": 147,
355
+ "Ķ": 148,
356
+ "ķ": 149,
357
+ "ĸ": 150,
358
+ "Ĺ": 151,
359
+ "ĺ": 152,
360
+ "Ļ": 153,
361
+ "ļ": 154,
362
+ "Ľ": 155,
363
+ "ľ": 156,
364
+ "Ŀ": 157,
365
+ "ŀ": 158,
366
+ "Ł": 159,
367
+ "ł": 160,
368
+ "¡": 161,
369
+ "¢": 162,
370
+ "£": 163,
371
+ "¤": 164,
372
+ "¥": 165,
373
+ "¦": 166,
374
+ "§": 167,
375
+ "¨": 168,
376
+ "©": 169,
377
+ "ª": 170,
378
+ "«": 171,
379
+ "¬": 172,
380
+ "Ń": 173,
381
+ "®": 174,
382
+ "¯": 175,
383
+ "°": 176,
384
+ "±": 177,
385
+ "²": 178,
386
+ "³": 179,
387
+ "´": 180,
388
+ "µ": 181,
389
+ "¶": 182,
390
+ "·": 183,
391
+ "¸": 184,
392
+ "¹": 185,
393
+ "º": 186,
394
+ "»": 187,
395
+ "¼": 188,
396
+ "½": 189,
397
+ "¾": 190,
398
+ "¿": 191,
399
+ "À": 192,
400
+ "Á": 193,
401
+ "Â": 194,
402
+ "Ã": 195,
403
+ "Ä": 196,
404
+ "Å": 197,
405
+ "Æ": 198,
406
+ "Ç": 199,
407
+ "È": 200,
408
+ "É": 201,
409
+ "Ê": 202,
410
+ "Ë": 203,
411
+ "Ì": 204,
412
+ "Í": 205,
413
+ "Î": 206,
414
+ "Ï": 207,
415
+ "Ð": 208,
416
+ "Ñ": 209,
417
+ "Ò": 210,
418
+ "Ó": 211,
419
+ "Ô": 212,
420
+ "Õ": 213,
421
+ "Ö": 214,
422
+ "×": 215,
423
+ "Ø": 216,
424
+ "Ù": 217,
425
+ "Ú": 218,
426
+ "Û": 219,
427
+ "Ü": 220,
428
+ "Ý": 221,
429
+ "Þ": 222,
430
+ "ß": 223,
431
+ "à": 224,
432
+ "á": 225,
433
+ "â": 226,
434
+ "ã": 227,
435
+ "ä": 228,
436
+ "å": 229,
437
+ "æ": 230,
438
+ "ç": 231,
439
+ "è": 232,
440
+ "é": 233,
441
+ "ê": 234,
442
+ "ë": 235,
443
+ "ì": 236,
444
+ "í": 237,
445
+ "î": 238,
446
+ "ï": 239,
447
+ "ð": 240,
448
+ "ñ": 241,
449
+ "ò": 242,
450
+ "ó": 243,
451
+ "ô": 244,
452
+ "õ": 245,
453
+ "ö": 246,
454
+ "÷": 247,
455
+ "ø": 248,
456
+ "ù": 249,
457
+ "ú": 250,
458
+ "û": 251,
459
+ "ü": 252,
460
+ "ý": 253,
461
+ "þ": 254,
462
+ "ÿ": 255,
463
+ "ĠĠ": 256,
464
+ "ĠĠĠĠ": 257,
465
+ "Ġt": 258,
466
+ "in": 259,
467
+ "er": 260,
468
+ "Ġa": 261,
469
+ "on": 262,
470
+ "en": 263,
471
+ "ä¸": 264,
472
+ "re": 265,
473
+ "he": 266,
474
+ "Ġs": 267,
475
+ "at": 268,
476
+ "ĠÐ": 269,
477
+ "or": 270,
478
+ "es": 271,
479
+ "Ġc": 272,
480
+ "о": 273,
481
+ "Ġd": 274,
482
+ "an": 275,
483
+ "Ġp": 276,
484
+ "ï¼": 277,
485
+ "it": 278,
486
+ "çļ": 279,
487
+ "ĠĠĠĠĠĠĠĠ": 280,
488
+ "çļĦ": 281,
489
+ "is": 282,
490
+ "Ġthe": 283,
491
+ "al": 284,
492
+ "ar": 285,
493
+ "е": 286,
494
+ "а": 287,
495
+ "ãĢ": 288,
496
+ "Ġo": 289,
497
+ "le": 290,
498
+ "Ġw": 291,
499
+ "Ġf": 292,
500
+ "äº": 293,
501
+ "Ġm": 294,
502
+ "ĠĠĠ": 295,
503
+ "и": 296,
504
+ "ou": 297,
505
+ "Ġb": 298,
506
+ "ÑĤ": 299,
507
+ "ï¼Į": 300,
508
+ "ic": 301,
509
+ "ing": 302,
510
+ "ion": 303,
511
+ "ed": 304,
512
+ "as": 305,
513
+ "nd": 306,
514
+ "н": 307,
515
+ "ro": 308,
516
+ "Ġin": 309,
517
+ "ent": 310,
518
+ "åı": 311,
519
+ "Ñģ": 312,
520
+ "ÑĢ": 313,
521
+ "æľ": 314,
522
+ "ä»": 315,
523
+ "el": 316,
524
+ "оÐ": 317,
525
+ "et": 318,
526
+ "Ġn": 319,
527
+ "Ġto": 320,
528
+ "st": 321,
529
+ "Ġl": 322,
530
+ "åħ": 323,
531
+ "âĢ": 324,
532
+ "Ġof": 325,
533
+ "ãĢĤ": 326,
534
+ "Ġh": 327,
535
+ "ct": 328,
536
+ "om": 329,
537
+ "å¤": 330,
538
+ "il": 331,
539
+ "Ġand": 332,
540
+ "åĪ": 333,
541
+ ".Ċ": 334,
542
+ "Ġth": 335,
543
+ "è¿": 336,
544
+ "å®": 337,
545
+ "ä½": 338,
546
+ "Ġre": 339,
547
+ "Ġe": 340,
548
+ "am": 341,
549
+ "id": 342,
550
+ "åIJ": 343,
551
+ "æĺ": 344,
552
+ "åľ": 345,
553
+ "ch": 346,
554
+ "çĶ": 347,
555
+ "im": 348,
556
+ "ut": 349,
557
+ "ur": 350,
558
+ ";Ċ": 351,
559
+ "ad": 352,
560
+ "л": 353,
561
+ "ol": 354,
562
+ "Ġg": 355,
563
+ "Ġv": 356,
564
+ "ç»": 357,
565
+ "Ġu": 358,
566
+ "ä¹": 359,
567
+ "Ġde": 360,
568
+ "æĪ": 361,
569
+ "åĬ": 362,
570
+ "ĠS": 363,
571
+ "ig": 364,
572
+ "ä¸Ģ": 365,
573
+ "se": 366,
574
+ "iv": 367,
575
+ "ãģ": 368,
576
+ "ot": 369,
577
+ "åĽ": 370,
578
+ "è¯": 371,
579
+ "æĺ¯": 372,
580
+ "æĸ": 373,
581
+ "ĠĠĠĠĠĠĠ": 374,
582
+ "ĠT": 375,
583
+ "ce": 376,
584
+ "Ġ=": 377,
585
+ "ĊĊ": 378,
586
+ "Ñĥ": 379,
587
+ "к": 380,
588
+ "å°": 381,
589
+ "ĠA": 382,
590
+ "ĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠĠ": 383,
591
+ "Ġy": 384,
592
+ "åŃ": 385,
593
+ "Ġ(": 386,
594
+ "ĠC": 387,
595
+ "æĹ": 388,
596
+ "ation": 389,
597
+ "qu": 390,
598
+ "ra": 391,
599
+ "ĠI": 392,
600
+ "ul": 393,
601
+ "--": 394,
602
+ "em": 395,
603
+ "ãĢģ": 396,
604
+ "os": 397,
605
+ "Ġis": 398,
606
+ "è®": 399,
607
+ "ow": 400,
608
+ "ä¼": 401,
609
+ "å¹": 402,
610
+ "od": 403,
611
+ "if": 404,
612
+ "est": 405,
613
+ "åį": 406,
614
+ "Ġst": 407,
615
+ "ag": 408,
616
+ "Ñı": 409,
617
+ "æĿ": 410,
618
+ "åľ¨": 411,
619
+ "Ġfor": 412,
620
+ "un": 413,
621
+ "Ġcon": 414,
622
+ "ä¸į": 415,
623
+ "ا": 416,
624
+ "æľī": 417,
625
+ "в": 418,
626
+ "ers": 419,
627
+ "ir": 420,
628
+ "м": 421,
629
+ "Ġ{": 422,
630
+ "Ñĭ": 423,
631
+ "Ġbe": 424,
632
+ "Ġ\"": 425,
633
+ "ay": 426,
634
+ "人": 427,
635
+ "æī": 428,
636
+ "å¼": 429,
637
+ "ab": 430,
638
+ "éĩ": 431,
639
+ "ter": 432,
640
+ "ĠM": 433,
641
+ "åº": 434,
642
+ "ĠP": 435,
643
+ "ate": 436,
644
+ "th": 437,
645
+ "äºĨ": 438,
646
+ "Ġon": 439,
647
+ "ly": 440,
648
+ "å·": 441,
649
+ "è¡": 442,
650
+ "us": 443,
651
+ "Ġп": 444,
652
+ "ac": 445,
653
+ "èĢ": 446,
654
+ "ĉĉ": 447,
655
+ "æķ": 448,
656
+ "åĨ": 449,
657
+ "ÑĮ": 450,
658
+ "åĩ": 451,
659
+ "ie": 452,
660
+ "å¯": 453,
661
+ "å¸": 454,
662
+ "Ġyou": 455,
663
+ "å¾": 456,
664
+ "Ġpro": 457,
665
+ "д": 458,
666
+ "pe": 459,
667
+ "ä¸Ń": 460,
668
+ "um": 461,
669
+ "åĮ": 462,
670
+ "ith": 463,
671
+ "ist": 464,
672
+ "éĢ": 465,
673
+ "为": 466,
674
+ "ÑģÑĤ": 467,
675
+ "æĢ": 468,
676
+ "Ġal": 469,
677
+ "oc": 470,
678
+ "é": 471,
679
+ "ĠB": 472,
680
+ "çĽ": 473,
681
+ "åĴ": 474,
682
+ "ĠD": 475,
683
+ "Ġthat": 476,
684
+ "ort": 477,
685
+ "ve": 478,
686
+ "res": 479,
687
+ "è§": 480,
688
+ "ĥ½": 481,
689
+ "Ġk": 482,
690
+ "ver": 483,
691
+ "ess": 484,
692
+ "te": 485,
693
+ "end": 486,
694
+ "Ġan": 487,
695
+ "Ġwh": 488,
696
+ "æł": 489,
697
+ "åĴĮ": 490,
698
+ "and": 491,
699
+ "Ġв": 492,
700
+ ",Ċ": 493,
701
+ "Ġ{Ċ": 494,
702
+ "ht": 495,
703
+ "ç§": 496,
704
+ ")Ċ": 497,
705
+ ");Ċ": 498,
706
+ "//": 499
707
+ },
708
+ "merges": [
709
+ [
710
+ "Ġ",
711
+ "Ġ"
712
+ ],
713
+ [
714
+ "Ġ",
715
+ "ĠĠĠ"
716
+ ],
717
+ [
718
+ "ĠĠ",
719
+ "ĠĠ"
720
+ ],
721
+ [
722
+ "ĠĠĠ",
723
+ "Ġ"
724
+ ],
725
+ [
726
+ "Ġ",
727
+ "t"
728
+ ],
729
+ [
730
+ "i",
731
+ "n"
732
+ ],
733
+ [
734
+ "e",
735
+ "r"
736
+ ],
737
+ [
738
+ "Ġ",
739
+ "a"
740
+ ],
741
+ [
742
+ "o",
743
+ "n"
744
+ ],
745
+ [
746
+ "e",
747
+ "n"
748
+ ],
749
+ [
750
+ "ä",
751
+ "¸"
752
+ ],
753
+ [
754
+ "r",
755
+ "e"
756
+ ],
757
+ [
758
+ "h",
759
+ "e"
760
+ ],
761
+ [
762
+ "Ġ",
763
+ "s"
764
+ ],
765
+ [
766
+ "a",
767
+ "t"
768
+ ],
769
+ [
770
+ "Ġ",
771
+ "Ð"
772
+ ],
773
+ [
774
+ "o",
775
+ "r"
776
+ ],
777
+ [
778
+ "e",
779
+ "s"
780
+ ],
781
+ [
782
+ "Ġ",
783
+ "c"
784
+ ],
785
+ [
786
+ "Ð",
787
+ "¾"
788
+ ],
789
+ [
790
+ "Ġ",
791
+ "d"
792
+ ],
793
+ [
794
+ "a",
795
+ "n"
796
+ ],
797
+ [
798
+ "Ġ",
799
+ "p"
800
+ ],
801
+ [
802
+ "ï",
803
+ "¼"
804
+ ],
805
+ [
806
+ "i",
807
+ "t"
808
+ ],
809
+ [
810
+ "ç",
811
+ "ļ"
812
+ ],
813
+ [
814
+ "Ġ",
815
+ "ĠĠĠĠĠĠĠ"
816
+ ],
817
+ [
818
+ "ĠĠĠĠ",
819
+ "ĠĠĠĠ"
820
+ ],
821
+ [
822
+ "ĠĠĠĠĠĠĠ",
823
+ "Ġ"
824
+ ],
825
+ [
826
+ "çļ",
827
+ "Ħ"
828
+ ],
829
+ [
830
+ "i",
831
+ "s"
832
+ ],
833
+ [
834
+ "Ġt",
835
+ "he"
836
+ ],
837
+ [
838
+ "Ġth",
839
+ "e"
840
+ ],
841
+ [
842
+ "a",
843
+ "l"
844
+ ],
845
+ [
846
+ "a",
847
+ "r"
848
+ ],
849
+ [
850
+ "Ð",
851
+ "µ"
852
+ ],
853
+ [
854
+ "Ð",
855
+ "°"
856
+ ],
857
+ [
858
+ "ã",
859
+ "Ģ"
860
+ ],
861
+ [
862
+ "Ġ",
863
+ "o"
864
+ ],
865
+ [
866
+ "l",
867
+ "e"
868
+ ],
869
+ [
870
+ "Ġ",
871
+ "w"
872
+ ],
873
+ [
874
+ "Ġ",
875
+ "f"
876
+ ],
877
+ [
878
+ "ä",
879
+ "º"
880
+ ],
881
+ [
882
+ "Ġ",
883
+ "m"
884
+ ],
885
+ [
886
+ "Ġ",
887
+ "ĠĠ"
888
+ ],
889
+ [
890
+ "ĠĠ",
891
+ "Ġ"
892
+ ],
893
+ [
894
+ "Ð",
895
+ "¸"
896
+ ],
897
+ [
898
+ "o",
899
+ "u"
900
+ ],
901
+ [
902
+ "Ġ",
903
+ "b"
904
+ ],
905
+ [
906
+ "Ñ",
907
+ "Ĥ"
908
+ ],
909
+ [
910
+ "ï¼",
911
+ "Į"
912
+ ],
913
+ [
914
+ "i",
915
+ "c"
916
+ ],
917
+ [
918
+ "in",
919
+ "g"
920
+ ],
921
+ [
922
+ "i",
923
+ "on"
924
+ ],
925
+ [
926
+ "e",
927
+ "d"
928
+ ],
929
+ [
930
+ "a",
931
+ "s"
932
+ ],
933
+ [
934
+ "n",
935
+ "d"
936
+ ],
937
+ [
938
+ "Ð",
939
+ "½"
940
+ ],
941
+ [
942
+ "r",
943
+ "o"
944
+ ],
945
+ [
946
+ "Ġ",
947
+ "in"
948
+ ],
949
+ [
950
+ "en",
951
+ "t"
952
+ ],
953
+ [
954
+ "å",
955
+ "ı"
956
+ ],
957
+ [
958
+ "Ñ",
959
+ "ģ"
960
+ ],
961
+ [
962
+ "Ñ",
963
+ "Ģ"
964
+ ],
965
+ [
966
+ "æ",
967
+ "ľ"
968
+ ],
969
+ [
970
+ "ä",
971
+ "»"
972
+ ],
973
+ [
974
+ "e",
975
+ "l"
976
+ ],
977
+ [
978
+ "о",
979
+ "Ð"
980
+ ],
981
+ [
982
+ "e",
983
+ "t"
984
+ ],
985
+ [
986
+ "Ġ",
987
+ "n"
988
+ ],
989
+ [
990
+ "Ġt",
991
+ "o"
992
+ ],
993
+ [
994
+ "s",
995
+ "t"
996
+ ],
997
+ [
998
+ "Ġ",
999
+ "l"
1000
+ ],
1001
+ [
1002
+ "å",
1003
+ "ħ"
1004
+ ],
1005
+ [
1006
+ "â",
1007
+ "Ģ"
1008
+ ],
1009
+ [
1010
+ "Ġo",
1011
+ "f"
1012
+ ],
1013
+ [
1014
+ "ãĢ",
1015
+ "Ĥ"
1016
+ ],
1017
+ [
1018
+ "Ġ",
1019
+ "h"
1020
+ ],
1021
+ [
1022
+ "c",
1023
+ "t"
1024
+ ],
1025
+ [
1026
+ "o",
1027
+ "m"
1028
+ ],
1029
+ [
1030
+ "å",
1031
+ "¤"
1032
+ ],
1033
+ [
1034
+ "i",
1035
+ "l"
1036
+ ],
1037
+ [
1038
+ "Ġ",
1039
+ "and"
1040
+ ],
1041
+ [
1042
+ "Ġa",
1043
+ "nd"
1044
+ ],
1045
+ [
1046
+ "Ġan",
1047
+ "d"
1048
+ ],
1049
+ [
1050
+ "å",
1051
+ "Ī"
1052
+ ],
1053
+ [
1054
+ ".",
1055
+ "Ċ"
1056
+ ],
1057
+ [
1058
+ "Ġ",
1059
+ "th"
1060
+ ],
1061
+ [
1062
+ "Ġt",
1063
+ "h"
1064
+ ],
1065
+ [
1066
+ "è",
1067
+ "¿"
1068
+ ],
1069
+ [
1070
+ "å",
1071
+ "®"
1072
+ ],
1073
+ [
1074
+ "ä",
1075
+ "½"
1076
+ ],
1077
+ [
1078
+ "Ġ",
1079
+ "re"
1080
+ ],
1081
+ [
1082
+ "Ġ",
1083
+ "e"
1084
+ ],
1085
+ [
1086
+ "a",
1087
+ "m"
1088
+ ],
1089
+ [
1090
+ "i",
1091
+ "d"
1092
+ ],
1093
+ [
1094
+ "å",
1095
+ "IJ"
1096
+ ],
1097
+ [
1098
+ "æ",
1099
+ "ĺ"
1100
+ ],
1101
+ [
1102
+ "å",
1103
+ "ľ"
1104
+ ],
1105
+ [
1106
+ "c",
1107
+ "h"
1108
+ ],
1109
+ [
1110
+ "ç",
1111
+ "Ķ"
1112
+ ],
1113
+ [
1114
+ "i",
1115
+ "m"
1116
+ ],
1117
+ [
1118
+ "u",
1119
+ "t"
1120
+ ],
1121
+ [
1122
+ "u",
1123
+ "r"
1124
+ ],
1125
+ [
1126
+ ";",
1127
+ "Ċ"
1128
+ ],
1129
+ [
1130
+ "a",
1131
+ "d"
1132
+ ],
1133
+ [
1134
+ "Ð",
1135
+ "»"
1136
+ ],
1137
+ [
1138
+ "o",
1139
+ "l"
1140
+ ],
1141
+ [
1142
+ "Ġ",
1143
+ "g"
1144
+ ],
1145
+ [
1146
+ "Ġ",
1147
+ "v"
1148
+ ],
1149
+ [
1150
+ "ç",
1151
+ "»"
1152
+ ],
1153
+ [
1154
+ "Ġ",
1155
+ "u"
1156
+ ],
1157
+ [
1158
+ "ä",
1159
+ "¹"
1160
+ ],
1161
+ [
1162
+ "Ġd",
1163
+ "e"
1164
+ ],
1165
+ [
1166
+ "æ",
1167
+ "Ī"
1168
+ ],
1169
+ [
1170
+ "å",
1171
+ "Ĭ"
1172
+ ],
1173
+ [
1174
+ "Ġ",
1175
+ "S"
1176
+ ],
1177
+ [
1178
+ "i",
1179
+ "g"
1180
+ ],
1181
+ [
1182
+ "ä¸",
1183
+ "Ģ"
1184
+ ],
1185
+ [
1186
+ "s",
1187
+ "e"
1188
+ ],
1189
+ [
1190
+ "i",
1191
+ "v"
1192
+ ],
1193
+ [
1194
+ "ã",
1195
+ "ģ"
1196
+ ],
1197
+ [
1198
+ "o",
1199
+ "t"
1200
+ ],
1201
+ [
1202
+ "å",
1203
+ "Ľ"
1204
+ ],
1205
+ [
1206
+ "è",
1207
+ "¯"
1208
+ ],
1209
+ [
1210
+ "æĺ",
1211
+ "¯"
1212
+ ],
1213
+ [
1214
+ "æ",
1215
+ "ĸ"
1216
+ ],
1217
+ [
1218
+ "ĠĠĠĠ",
1219
+ "ĠĠĠ"
1220
+ ],
1221
+ [
1222
+ "ĠĠĠ",
1223
+ "ĠĠĠĠ"
1224
+ ],
1225
+ [
1226
+ "Ġ",
1227
+ "T"
1228
+ ],
1229
+ [
1230
+ "c",
1231
+ "e"
1232
+ ],
1233
+ [
1234
+ "Ġ",
1235
+ "="
1236
+ ],
1237
+ [
1238
+ "Ċ",
1239
+ "Ċ"
1240
+ ],
1241
+ [
1242
+ "Ñ",
1243
+ "ĥ"
1244
+ ],
1245
+ [
1246
+ "Ð",
1247
+ "º"
1248
+ ],
1249
+ [
1250
+ "å",
1251
+ "°"
1252
+ ],
1253
+ [
1254
+ "Ġ",
1255
+ "A"
1256
+ ],
1257
+ [
1258
+ "ĠĠĠĠĠĠĠĠ",
1259
+ "ĠĠĠĠĠĠĠĠ"
1260
+ ],
1261
+ [
1262
+ "Ġ",
1263
+ "y"
1264
+ ],
1265
+ [
1266
+ "å",
1267
+ "Ń"
1268
+ ],
1269
+ [
1270
+ "Ġ",
1271
+ "("
1272
+ ],
1273
+ [
1274
+ "Ġ",
1275
+ "C"
1276
+ ],
1277
+ [
1278
+ "æ",
1279
+ "Ĺ"
1280
+ ],
1281
+ [
1282
+ "at",
1283
+ "ion"
1284
+ ],
1285
+ [
1286
+ "q",
1287
+ "u"
1288
+ ],
1289
+ [
1290
+ "r",
1291
+ "a"
1292
+ ],
1293
+ [
1294
+ "Ġ",
1295
+ "I"
1296
+ ],
1297
+ [
1298
+ "u",
1299
+ "l"
1300
+ ],
1301
+ [
1302
+ "-",
1303
+ "-"
1304
+ ],
1305
+ [
1306
+ "e",
1307
+ "m"
1308
+ ],
1309
+ [
1310
+ "ãĢ",
1311
+ "ģ"
1312
+ ],
1313
+ [
1314
+ "o",
1315
+ "s"
1316
+ ],
1317
+ [
1318
+ "Ġ",
1319
+ "is"
1320
+ ],
1321
+ [
1322
+ "è",
1323
+ "®"
1324
+ ],
1325
+ [
1326
+ "o",
1327
+ "w"
1328
+ ],
1329
+ [
1330
+ "ä",
1331
+ "¼"
1332
+ ],
1333
+ [
1334
+ "å",
1335
+ "¹"
1336
+ ],
1337
+ [
1338
+ "o",
1339
+ "d"
1340
+ ],
1341
+ [
1342
+ "i",
1343
+ "f"
1344
+ ],
1345
+ [
1346
+ "e",
1347
+ "st"
1348
+ ],
1349
+ [
1350
+ "es",
1351
+ "t"
1352
+ ],
1353
+ [
1354
+ "å",
1355
+ "į"
1356
+ ],
1357
+ [
1358
+ "Ġ",
1359
+ "st"
1360
+ ],
1361
+ [
1362
+ "Ġs",
1363
+ "t"
1364
+ ],
1365
+ [
1366
+ "a",
1367
+ "g"
1368
+ ],
1369
+ [
1370
+ "Ñ",
1371
+ "ı"
1372
+ ],
1373
+ [
1374
+ "æ",
1375
+ "Ŀ"
1376
+ ],
1377
+ [
1378
+ "åľ",
1379
+ "¨"
1380
+ ],
1381
+ [
1382
+ "Ġf",
1383
+ "or"
1384
+ ],
1385
+ [
1386
+ "u",
1387
+ "n"
1388
+ ],
1389
+ [
1390
+ "Ġc",
1391
+ "on"
1392
+ ],
1393
+ [
1394
+ "ä¸",
1395
+ "į"
1396
+ ],
1397
+ [
1398
+ "Ø",
1399
+ "§"
1400
+ ],
1401
+ [
1402
+ "æľ",
1403
+ "ī"
1404
+ ],
1405
+ [
1406
+ "Ð",
1407
+ "²"
1408
+ ],
1409
+ [
1410
+ "er",
1411
+ "s"
1412
+ ],
1413
+ [
1414
+ "i",
1415
+ "r"
1416
+ ],
1417
+ [
1418
+ "Ð",
1419
+ "¼"
1420
+ ],
1421
+ [
1422
+ "Ġ",
1423
+ "{"
1424
+ ],
1425
+ [
1426
+ "Ñ",
1427
+ "ĭ"
1428
+ ],
1429
+ [
1430
+ "Ġb",
1431
+ "e"
1432
+ ],
1433
+ [
1434
+ "Ġ",
1435
+ "\""
1436
+ ],
1437
+ [
1438
+ "a",
1439
+ "y"
1440
+ ],
1441
+ [
1442
+ "äº",
1443
+ "º"
1444
+ ],
1445
+ [
1446
+ "æ",
1447
+ "ī"
1448
+ ],
1449
+ [
1450
+ "å",
1451
+ "¼"
1452
+ ],
1453
+ [
1454
+ "a",
1455
+ "b"
1456
+ ],
1457
+ [
1458
+ "é",
1459
+ "ĩ"
1460
+ ],
1461
+ [
1462
+ "t",
1463
+ "er"
1464
+ ],
1465
+ [
1466
+ "te",
1467
+ "r"
1468
+ ],
1469
+ [
1470
+ "Ġ",
1471
+ "M"
1472
+ ],
1473
+ [
1474
+ "å",
1475
+ "º"
1476
+ ],
1477
+ [
1478
+ "Ġ",
1479
+ "P"
1480
+ ],
1481
+ [
1482
+ "a",
1483
+ "te"
1484
+ ],
1485
+ [
1486
+ "at",
1487
+ "e"
1488
+ ],
1489
+ [
1490
+ "t",
1491
+ "h"
1492
+ ],
1493
+ [
1494
+ "äº",
1495
+ "Ĩ"
1496
+ ],
1497
+ [
1498
+ "Ġ",
1499
+ "on"
1500
+ ],
1501
+ [
1502
+ "Ġo",
1503
+ "n"
1504
+ ],
1505
+ [
1506
+ "l",
1507
+ "y"
1508
+ ],
1509
+ [
1510
+ "å",
1511
+ "·"
1512
+ ],
1513
+ [
1514
+ "è",
1515
+ "¡"
1516
+ ],
1517
+ [
1518
+ "u",
1519
+ "s"
1520
+ ],
1521
+ [
1522
+ "ĠÐ",
1523
+ "¿"
1524
+ ],
1525
+ [
1526
+ "a",
1527
+ "c"
1528
+ ],
1529
+ [
1530
+ "è",
1531
+ "Ģ"
1532
+ ],
1533
+ [
1534
+ "ĉ",
1535
+ "ĉ"
1536
+ ],
1537
+ [
1538
+ "æ",
1539
+ "ķ"
1540
+ ],
1541
+ [
1542
+ "å",
1543
+ "Ĩ"
1544
+ ],
1545
+ [
1546
+ "Ñ",
1547
+ "Į"
1548
+ ],
1549
+ [
1550
+ "å",
1551
+ "ĩ"
1552
+ ],
1553
+ [
1554
+ "i",
1555
+ "e"
1556
+ ],
1557
+ [
1558
+ "å",
1559
+ "¯"
1560
+ ],
1561
+ [
1562
+ "å",
1563
+ "¸"
1564
+ ],
1565
+ [
1566
+ "Ġy",
1567
+ "ou"
1568
+ ],
1569
+ [
1570
+ "å",
1571
+ "¾"
1572
+ ],
1573
+ [
1574
+ "Ġp",
1575
+ "ro"
1576
+ ],
1577
+ [
1578
+ "Ð",
1579
+ "´"
1580
+ ],
1581
+ [
1582
+ "p",
1583
+ "e"
1584
+ ],
1585
+ [
1586
+ "ä¸",
1587
+ "Ń"
1588
+ ],
1589
+ [
1590
+ "u",
1591
+ "m"
1592
+ ],
1593
+ [
1594
+ "å",
1595
+ "Į"
1596
+ ],
1597
+ [
1598
+ "i",
1599
+ "th"
1600
+ ],
1601
+ [
1602
+ "it",
1603
+ "h"
1604
+ ],
1605
+ [
1606
+ "i",
1607
+ "st"
1608
+ ],
1609
+ [
1610
+ "is",
1611
+ "t"
1612
+ ],
1613
+ [
1614
+ "é",
1615
+ "Ģ"
1616
+ ],
1617
+ [
1618
+ "ä¸",
1619
+ "º"
1620
+ ],
1621
+ [
1622
+ "Ñģ",
1623
+ "ÑĤ"
1624
+ ],
1625
+ [
1626
+ "æ",
1627
+ "Ģ"
1628
+ ],
1629
+ [
1630
+ "Ġ",
1631
+ "al"
1632
+ ],
1633
+ [
1634
+ "Ġa",
1635
+ "l"
1636
+ ],
1637
+ [
1638
+ "o",
1639
+ "c"
1640
+ ],
1641
+ [
1642
+ "Ã",
1643
+ "©"
1644
+ ],
1645
+ [
1646
+ "Ġ",
1647
+ "B"
1648
+ ],
1649
+ [
1650
+ "ç",
1651
+ "Ľ"
1652
+ ],
1653
+ [
1654
+ "å",
1655
+ "Ĵ"
1656
+ ],
1657
+ [
1658
+ "Ġ",
1659
+ "D"
1660
+ ],
1661
+ [
1662
+ "Ġth",
1663
+ "at"
1664
+ ],
1665
+ [
1666
+ "or",
1667
+ "t"
1668
+ ],
1669
+ [
1670
+ "v",
1671
+ "e"
1672
+ ],
1673
+ [
1674
+ "r",
1675
+ "es"
1676
+ ],
1677
+ [
1678
+ "re",
1679
+ "s"
1680
+ ],
1681
+ [
1682
+ "è",
1683
+ "§"
1684
+ ],
1685
+ [
1686
+ "ĥ",
1687
+ "½"
1688
+ ],
1689
+ [
1690
+ "Ġ",
1691
+ "k"
1692
+ ],
1693
+ [
1694
+ "v",
1695
+ "er"
1696
+ ],
1697
+ [
1698
+ "ve",
1699
+ "r"
1700
+ ],
1701
+ [
1702
+ "es",
1703
+ "s"
1704
+ ],
1705
+ [
1706
+ "t",
1707
+ "e"
1708
+ ],
1709
+ [
1710
+ "e",
1711
+ "nd"
1712
+ ],
1713
+ [
1714
+ "en",
1715
+ "d"
1716
+ ],
1717
+ [
1718
+ "Ġ",
1719
+ "an"
1720
+ ],
1721
+ [
1722
+ "Ġa",
1723
+ "n"
1724
+ ],
1725
+ [
1726
+ "Ġw",
1727
+ "h"
1728
+ ],
1729
+ [
1730
+ "æ",
1731
+ "ł"
1732
+ ],
1733
+ [
1734
+ "åĴ",
1735
+ "Į"
1736
+ ],
1737
+ [
1738
+ "a",
1739
+ "nd"
1740
+ ],
1741
+ [
1742
+ "an",
1743
+ "d"
1744
+ ],
1745
+ [
1746
+ "Ġ",
1747
+ "в"
1748
+ ],
1749
+ [
1750
+ "ĠÐ",
1751
+ "²"
1752
+ ],
1753
+ [
1754
+ ",",
1755
+ "Ċ"
1756
+ ],
1757
+ [
1758
+ "Ġ{",
1759
+ "Ċ"
1760
+ ],
1761
+ [
1762
+ "h",
1763
+ "t"
1764
+ ],
1765
+ [
1766
+ "ç",
1767
+ "§"
1768
+ ],
1769
+ [
1770
+ ")",
1771
+ "Ċ"
1772
+ ],
1773
+ [
1774
+ ")",
1775
+ ";Ċ"
1776
+ ],
1777
+ [
1778
+ "/",
1779
+ "/"
1780
+ ]
1781
+ ]
1782
+ }
1783
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,175 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "clean_up_tokenization_spaces": false,
4
+ "do_lower_case": false,
5
+ "eos_token": "<|endoftext|>",
6
+ "extra_special_tokens": [
7
+ "<|endoftext|>",
8
+ "[MASK]",
9
+ "[gMASK]",
10
+ "[sMASK]",
11
+ "<sop>",
12
+ "<eop>",
13
+ "<|system|>",
14
+ "<|user|>",
15
+ "<|assistant|>",
16
+ "<|observation|>",
17
+ "<|begin_of_image|>",
18
+ "<|end_of_image|>",
19
+ "<|begin_of_video|>",
20
+ "<|end_of_video|>",
21
+ "<|pad|>",
22
+ "<|begin_of_audio|>",
23
+ "<|end_of_audio|>"
24
+ ],
25
+ "is_local": false,
26
+ "model_input_names": [
27
+ "input_ids",
28
+ "attention_mask"
29
+ ],
30
+ "model_max_length": 65536,
31
+ "model_specific_special_tokens": {},
32
+ "pad_token": "<|endoftext|>",
33
+ "padding_side": "left",
34
+ "processor_class": "GlmAsrProcessor",
35
+ "remove_space": false,
36
+ "tokenizer_class": "TokenizersBackend",
37
+ "added_tokens_decoder": {
38
+ "500": {
39
+ "content": "<|endoftext|>",
40
+ "single_word": false,
41
+ "lstrip": false,
42
+ "rstrip": false,
43
+ "normalized": false,
44
+ "special": true
45
+ },
46
+ "501": {
47
+ "content": "[MASK]",
48
+ "single_word": false,
49
+ "lstrip": false,
50
+ "rstrip": false,
51
+ "normalized": false,
52
+ "special": true
53
+ },
54
+ "502": {
55
+ "content": "[gMASK]",
56
+ "single_word": false,
57
+ "lstrip": false,
58
+ "rstrip": false,
59
+ "normalized": false,
60
+ "special": true
61
+ },
62
+ "503": {
63
+ "content": "[sMASK]",
64
+ "single_word": false,
65
+ "lstrip": false,
66
+ "rstrip": false,
67
+ "normalized": false,
68
+ "special": true
69
+ },
70
+ "504": {
71
+ "content": "<sop>",
72
+ "single_word": false,
73
+ "lstrip": false,
74
+ "rstrip": false,
75
+ "normalized": false,
76
+ "special": true
77
+ },
78
+ "505": {
79
+ "content": "<eop>",
80
+ "single_word": false,
81
+ "lstrip": false,
82
+ "rstrip": false,
83
+ "normalized": false,
84
+ "special": true
85
+ },
86
+ "506": {
87
+ "content": "<|system|>",
88
+ "single_word": false,
89
+ "lstrip": false,
90
+ "rstrip": false,
91
+ "normalized": false,
92
+ "special": true
93
+ },
94
+ "507": {
95
+ "content": "<|user|>",
96
+ "single_word": false,
97
+ "lstrip": false,
98
+ "rstrip": false,
99
+ "normalized": false,
100
+ "special": true
101
+ },
102
+ "508": {
103
+ "content": "<|assistant|>",
104
+ "single_word": false,
105
+ "lstrip": false,
106
+ "rstrip": false,
107
+ "normalized": false,
108
+ "special": true
109
+ },
110
+ "509": {
111
+ "content": "<|observation|>",
112
+ "single_word": false,
113
+ "lstrip": false,
114
+ "rstrip": false,
115
+ "normalized": false,
116
+ "special": true
117
+ },
118
+ "510": {
119
+ "content": "<|begin_of_image|>",
120
+ "single_word": false,
121
+ "lstrip": false,
122
+ "rstrip": false,
123
+ "normalized": false,
124
+ "special": true
125
+ },
126
+ "511": {
127
+ "content": "<|end_of_image|>",
128
+ "single_word": false,
129
+ "lstrip": false,
130
+ "rstrip": false,
131
+ "normalized": false,
132
+ "special": true
133
+ },
134
+ "512": {
135
+ "content": "<|begin_of_video|>",
136
+ "single_word": false,
137
+ "lstrip": false,
138
+ "rstrip": false,
139
+ "normalized": false,
140
+ "special": true
141
+ },
142
+ "513": {
143
+ "content": "<|end_of_video|>",
144
+ "single_word": false,
145
+ "lstrip": false,
146
+ "rstrip": false,
147
+ "normalized": false,
148
+ "special": true
149
+ },
150
+ "514": {
151
+ "content": "<|pad|>",
152
+ "single_word": false,
153
+ "lstrip": false,
154
+ "rstrip": false,
155
+ "normalized": false,
156
+ "special": true
157
+ },
158
+ "515": {
159
+ "content": "<|begin_of_audio|>",
160
+ "single_word": false,
161
+ "lstrip": false,
162
+ "rstrip": false,
163
+ "normalized": false,
164
+ "special": true
165
+ },
166
+ "516": {
167
+ "content": "<|end_of_audio|>",
168
+ "single_word": false,
169
+ "lstrip": false,
170
+ "rstrip": false,
171
+ "normalized": false,
172
+ "special": true
173
+ }
174
+ }
175
+ }