yesj1234 commited on
Commit
569435c
1 Parent(s): 459813d

Upload folder using huggingface_hub

Browse files
README.md ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: facebook/wav2vec2-large-xlsr-53
4
+ tags:
5
+ - automatic-speech-recognition
6
+ - ./sample_speech.py
7
+ - generated_from_trainer
8
+ model-index:
9
+ - name: ko-xlsr
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # ko-xlsr
17
+
18
+ This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the ./SAMPLE_SPEECH.PY - NA dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 0.5156
21
+ - Cer: 0.1228
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 0.0003
41
+ - train_batch_size: 4
42
+ - eval_batch_size: 4
43
+ - seed: 42
44
+ - distributed_type: multi-GPU
45
+ - num_devices: 4
46
+ - gradient_accumulation_steps: 8
47
+ - total_train_batch_size: 128
48
+ - total_eval_batch_size: 16
49
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
+ - lr_scheduler_type: linear
51
+ - lr_scheduler_warmup_steps: 500
52
+ - num_epochs: 40
53
+
54
+ ### Training results
55
+
56
+ | Training Loss | Epoch | Step | Validation Loss | Cer |
57
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
58
+ | 1.778 | 3.31 | 1000 | 1.2773 | 0.3050 |
59
+ | 1.1037 | 6.63 | 2000 | 0.7716 | 0.1888 |
60
+ | 0.9529 | 9.94 | 3000 | 0.6726 | 0.1659 |
61
+ | 0.8424 | 13.26 | 4000 | 0.6138 | 0.1512 |
62
+ | 0.767 | 16.57 | 5000 | 0.5885 | 0.1433 |
63
+ | 0.7201 | 19.88 | 6000 | 0.5682 | 0.1378 |
64
+ | 0.664 | 23.2 | 7000 | 0.5583 | 0.1333 |
65
+ | 0.6296 | 26.51 | 8000 | 0.5416 | 0.1298 |
66
+ | 0.6021 | 29.83 | 9000 | 0.5377 | 0.1272 |
67
+ | 0.568 | 33.14 | 10000 | 0.5241 | 0.1246 |
68
+ | 0.5519 | 36.45 | 11000 | 0.5184 | 0.1228 |
69
+ | 0.5395 | 39.77 | 12000 | 0.5156 | 0.1227 |
70
+
71
+
72
+ ### Framework versions
73
+
74
+ - Transformers 4.33.0
75
+ - Pytorch 2.0.1+cu117
76
+ - Datasets 2.14.4
77
+ - Tokenizers 0.13.3
added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "</s>": 1431,
3
+ "<s>": 1430
4
+ }
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 39.9,
3
+ "eval_cer": 0.12280162938057675,
4
+ "eval_loss": 0.51564621925354,
5
+ "eval_runtime": 141.8188,
6
+ "eval_samples": 4839,
7
+ "eval_samples_per_second": 34.121,
8
+ "eval_steps_per_second": 2.137,
9
+ "train_loss": 1.365207207163307,
10
+ "train_runtime": 27512.8703,
11
+ "train_samples": 38612,
12
+ "train_samples_per_second": 56.137,
13
+ "train_steps_per_second": 0.438
14
+ }
checkpoint-12000/added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "</s>": 1431,
3
+ "<s>": 1430
4
+ }
checkpoint-12000/config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForCTC"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 768,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "mean",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.1,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "gradient_checkpointing": false,
57
+ "hidden_act": "gelu",
58
+ "hidden_dropout": 0.1,
59
+ "hidden_size": 1024,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "layer_norm_eps": 1e-05,
63
+ "layerdrop": 0.1,
64
+ "mask_channel_length": 10,
65
+ "mask_channel_min_space": 1,
66
+ "mask_channel_other": 0.0,
67
+ "mask_channel_prob": 0.0,
68
+ "mask_channel_selection": "static",
69
+ "mask_feature_length": 64,
70
+ "mask_feature_min_masks": 0,
71
+ "mask_feature_prob": 0.0,
72
+ "mask_time_length": 10,
73
+ "mask_time_min_masks": 2,
74
+ "mask_time_min_space": 1,
75
+ "mask_time_other": 0.0,
76
+ "mask_time_prob": 0.3,
77
+ "mask_time_selection": "static",
78
+ "model_type": "wav2vec2",
79
+ "num_adapter_layers": 3,
80
+ "num_attention_heads": 16,
81
+ "num_codevector_groups": 2,
82
+ "num_codevectors_per_group": 320,
83
+ "num_conv_pos_embedding_groups": 16,
84
+ "num_conv_pos_embeddings": 128,
85
+ "num_feat_extract_layers": 7,
86
+ "num_hidden_layers": 24,
87
+ "num_negatives": 100,
88
+ "output_hidden_size": 1024,
89
+ "pad_token_id": 1429,
90
+ "proj_codevector_dim": 768,
91
+ "tdnn_dilation": [
92
+ 1,
93
+ 2,
94
+ 3,
95
+ 1,
96
+ 1
97
+ ],
98
+ "tdnn_dim": [
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 512,
103
+ 1500
104
+ ],
105
+ "tdnn_kernel": [
106
+ 5,
107
+ 3,
108
+ 3,
109
+ 1,
110
+ 1
111
+ ],
112
+ "torch_dtype": "float32",
113
+ "transformers_version": "4.33.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 1432,
116
+ "xvector_output_dim": 512
117
+ }
checkpoint-12000/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c87fd25c8943ab89a434e0fb03144d7bd5891ad194a363dde2d0fc47d9b9f37
3
+ size 2501901697
checkpoint-12000/preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
checkpoint-12000/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a39f413b63d4c11c976546d13301948eea94e33f6af0550035981aac625a602
3
+ size 1267773037
checkpoint-12000/rng_state_0.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad981acca3d2fda15529e54de3b68cbb0acc4f83d3297151940d844ff4d65de5
3
+ size 17719
checkpoint-12000/rng_state_1.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7d1348634df138e00b51720d802c8c59680dcc6c69aaac705679c9540204fa1
3
+ size 17655
checkpoint-12000/rng_state_2.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e5e504393cf0594ad6e2898f4508e96be88a4688a0f8a726f1ae36a538cb366
3
+ size 17655
checkpoint-12000/rng_state_3.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cf98284e3e77ff32a8afc8cc5972950a802df5a0dd4849a1a545dfe12115964b
3
+ size 17719
checkpoint-12000/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c48e4d7456f8edb0a67340795bda06f87a28d576080a962121a5c7f1f5c76e45
3
+ size 627
checkpoint-12000/special_tokens_map.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": true,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ {
11
+ "content": "</s>",
12
+ "lstrip": false,
13
+ "normalized": true,
14
+ "rstrip": false,
15
+ "single_word": false
16
+ },
17
+ {
18
+ "content": "<s>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ {
25
+ "content": "</s>",
26
+ "lstrip": false,
27
+ "normalized": true,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ ],
32
+ "bos_token": "<s>",
33
+ "eos_token": "</s>",
34
+ "pad_token": "[PAD]",
35
+ "unk_token": "[UNK]"
36
+ }
checkpoint-12000/tokenizer_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "clean_up_tokenization_spaces": true,
4
+ "config": null,
5
+ "do_lower_case": false,
6
+ "eos_token": "</s>",
7
+ "model_max_length": 1000000000000000019884624838656,
8
+ "pad_token": "[PAD]",
9
+ "processor_class": "Wav2Vec2Processor",
10
+ "replace_word_delimiter_char": " ",
11
+ "target_lang": null,
12
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
13
+ "tokenizer_type": "wav2vec2",
14
+ "trust_remote_code": false,
15
+ "unk_token": "[UNK]",
16
+ "word_delimiter_token": "|"
17
+ }
checkpoint-12000/trainer_state.json ADDED
@@ -0,0 +1,847 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 39.768019884009945,
5
+ "eval_steps": 1000,
6
+ "global_step": 12000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.33,
13
+ "learning_rate": 5.6399999999999995e-05,
14
+ "loss": 37.3417,
15
+ "step": 100
16
+ },
17
+ {
18
+ "epoch": 0.66,
19
+ "learning_rate": 0.0001164,
20
+ "loss": 12.5064,
21
+ "step": 200
22
+ },
23
+ {
24
+ "epoch": 0.99,
25
+ "learning_rate": 0.00017639999999999998,
26
+ "loss": 4.823,
27
+ "step": 300
28
+ },
29
+ {
30
+ "epoch": 1.33,
31
+ "learning_rate": 0.0002364,
32
+ "loss": 4.7159,
33
+ "step": 400
34
+ },
35
+ {
36
+ "epoch": 1.66,
37
+ "learning_rate": 0.0002964,
38
+ "loss": 4.6828,
39
+ "step": 500
40
+ },
41
+ {
42
+ "epoch": 1.99,
43
+ "learning_rate": 0.0002975563258232235,
44
+ "loss": 4.5391,
45
+ "step": 600
46
+ },
47
+ {
48
+ "epoch": 2.32,
49
+ "learning_rate": 0.0002949566724436741,
50
+ "loss": 4.3365,
51
+ "step": 700
52
+ },
53
+ {
54
+ "epoch": 2.65,
55
+ "learning_rate": 0.0002923570190641247,
56
+ "loss": 3.2316,
57
+ "step": 800
58
+ },
59
+ {
60
+ "epoch": 2.98,
61
+ "learning_rate": 0.00028975736568457536,
62
+ "loss": 2.1607,
63
+ "step": 900
64
+ },
65
+ {
66
+ "epoch": 3.31,
67
+ "learning_rate": 0.00028715771230502596,
68
+ "loss": 1.778,
69
+ "step": 1000
70
+ },
71
+ {
72
+ "epoch": 3.31,
73
+ "eval_cer": 0.3050481701797491,
74
+ "eval_loss": 1.2772819995880127,
75
+ "eval_runtime": 143.2917,
76
+ "eval_samples_per_second": 33.77,
77
+ "eval_steps_per_second": 2.115,
78
+ "step": 1000
79
+ },
80
+ {
81
+ "epoch": 3.65,
82
+ "learning_rate": 0.00028455805892547655,
83
+ "loss": 1.6287,
84
+ "step": 1100
85
+ },
86
+ {
87
+ "epoch": 3.98,
88
+ "learning_rate": 0.0002819584055459272,
89
+ "loss": 1.5046,
90
+ "step": 1200
91
+ },
92
+ {
93
+ "epoch": 4.31,
94
+ "learning_rate": 0.0002793587521663778,
95
+ "loss": 1.4046,
96
+ "step": 1300
97
+ },
98
+ {
99
+ "epoch": 4.64,
100
+ "learning_rate": 0.0002767590987868284,
101
+ "loss": 1.3378,
102
+ "step": 1400
103
+ },
104
+ {
105
+ "epoch": 4.97,
106
+ "learning_rate": 0.000274159445407279,
107
+ "loss": 1.2868,
108
+ "step": 1500
109
+ },
110
+ {
111
+ "epoch": 5.3,
112
+ "learning_rate": 0.0002715597920277296,
113
+ "loss": 1.2376,
114
+ "step": 1600
115
+ },
116
+ {
117
+ "epoch": 5.63,
118
+ "learning_rate": 0.00026896013864818023,
119
+ "loss": 1.2126,
120
+ "step": 1700
121
+ },
122
+ {
123
+ "epoch": 5.97,
124
+ "learning_rate": 0.0002663604852686308,
125
+ "loss": 1.1781,
126
+ "step": 1800
127
+ },
128
+ {
129
+ "epoch": 6.3,
130
+ "learning_rate": 0.0002637608318890814,
131
+ "loss": 1.1426,
132
+ "step": 1900
133
+ },
134
+ {
135
+ "epoch": 6.63,
136
+ "learning_rate": 0.00026116117850953207,
137
+ "loss": 1.1037,
138
+ "step": 2000
139
+ },
140
+ {
141
+ "epoch": 6.63,
142
+ "eval_cer": 0.18878507694297167,
143
+ "eval_loss": 0.7715919017791748,
144
+ "eval_runtime": 143.3347,
145
+ "eval_samples_per_second": 33.76,
146
+ "eval_steps_per_second": 2.114,
147
+ "step": 2000
148
+ },
149
+ {
150
+ "epoch": 6.96,
151
+ "learning_rate": 0.00025856152512998266,
152
+ "loss": 1.0971,
153
+ "step": 2100
154
+ },
155
+ {
156
+ "epoch": 7.29,
157
+ "learning_rate": 0.00025596187175043326,
158
+ "loss": 1.0693,
159
+ "step": 2200
160
+ },
161
+ {
162
+ "epoch": 7.62,
163
+ "learning_rate": 0.00025336221837088385,
164
+ "loss": 1.0398,
165
+ "step": 2300
166
+ },
167
+ {
168
+ "epoch": 7.95,
169
+ "learning_rate": 0.00025076256499133445,
170
+ "loss": 1.0282,
171
+ "step": 2400
172
+ },
173
+ {
174
+ "epoch": 8.29,
175
+ "learning_rate": 0.00024816291161178504,
176
+ "loss": 0.9997,
177
+ "step": 2500
178
+ },
179
+ {
180
+ "epoch": 8.62,
181
+ "learning_rate": 0.0002455632582322357,
182
+ "loss": 0.995,
183
+ "step": 2600
184
+ },
185
+ {
186
+ "epoch": 8.95,
187
+ "learning_rate": 0.00024296360485268629,
188
+ "loss": 0.9721,
189
+ "step": 2700
190
+ },
191
+ {
192
+ "epoch": 9.28,
193
+ "learning_rate": 0.00024036395147313688,
194
+ "loss": 0.9639,
195
+ "step": 2800
196
+ },
197
+ {
198
+ "epoch": 9.61,
199
+ "learning_rate": 0.0002377642980935875,
200
+ "loss": 0.9543,
201
+ "step": 2900
202
+ },
203
+ {
204
+ "epoch": 9.94,
205
+ "learning_rate": 0.0002351646447140381,
206
+ "loss": 0.9529,
207
+ "step": 3000
208
+ },
209
+ {
210
+ "epoch": 9.94,
211
+ "eval_cer": 0.16591232380706064,
212
+ "eval_loss": 0.6726131439208984,
213
+ "eval_runtime": 144.4701,
214
+ "eval_samples_per_second": 33.495,
215
+ "eval_steps_per_second": 2.097,
216
+ "step": 3000
217
+ },
218
+ {
219
+ "epoch": 10.27,
220
+ "learning_rate": 0.0002325649913344887,
221
+ "loss": 0.9281,
222
+ "step": 3100
223
+ },
224
+ {
225
+ "epoch": 10.6,
226
+ "learning_rate": 0.00022996533795493934,
227
+ "loss": 0.9274,
228
+ "step": 3200
229
+ },
230
+ {
231
+ "epoch": 10.94,
232
+ "learning_rate": 0.00022736568457538994,
233
+ "loss": 0.9149,
234
+ "step": 3300
235
+ },
236
+ {
237
+ "epoch": 11.27,
238
+ "learning_rate": 0.00022476603119584053,
239
+ "loss": 0.9036,
240
+ "step": 3400
241
+ },
242
+ {
243
+ "epoch": 11.6,
244
+ "learning_rate": 0.00022216637781629115,
245
+ "loss": 0.8838,
246
+ "step": 3500
247
+ },
248
+ {
249
+ "epoch": 11.93,
250
+ "learning_rate": 0.00021956672443674175,
251
+ "loss": 0.8872,
252
+ "step": 3600
253
+ },
254
+ {
255
+ "epoch": 12.26,
256
+ "learning_rate": 0.00021696707105719234,
257
+ "loss": 0.8661,
258
+ "step": 3700
259
+ },
260
+ {
261
+ "epoch": 12.59,
262
+ "learning_rate": 0.00021436741767764296,
263
+ "loss": 0.8529,
264
+ "step": 3800
265
+ },
266
+ {
267
+ "epoch": 12.92,
268
+ "learning_rate": 0.00021176776429809356,
269
+ "loss": 0.8566,
270
+ "step": 3900
271
+ },
272
+ {
273
+ "epoch": 13.26,
274
+ "learning_rate": 0.00020916811091854415,
275
+ "loss": 0.8424,
276
+ "step": 4000
277
+ },
278
+ {
279
+ "epoch": 13.26,
280
+ "eval_cer": 0.1511541445751972,
281
+ "eval_loss": 0.6138216257095337,
282
+ "eval_runtime": 143.2958,
283
+ "eval_samples_per_second": 33.769,
284
+ "eval_steps_per_second": 2.115,
285
+ "step": 4000
286
+ },
287
+ {
288
+ "epoch": 13.59,
289
+ "learning_rate": 0.0002065684575389948,
290
+ "loss": 0.8403,
291
+ "step": 4100
292
+ },
293
+ {
294
+ "epoch": 13.92,
295
+ "learning_rate": 0.0002039688041594454,
296
+ "loss": 0.836,
297
+ "step": 4200
298
+ },
299
+ {
300
+ "epoch": 14.25,
301
+ "learning_rate": 0.000201369150779896,
302
+ "loss": 0.8119,
303
+ "step": 4300
304
+ },
305
+ {
306
+ "epoch": 14.58,
307
+ "learning_rate": 0.0001987694974003466,
308
+ "loss": 0.8298,
309
+ "step": 4400
310
+ },
311
+ {
312
+ "epoch": 14.91,
313
+ "learning_rate": 0.0001961698440207972,
314
+ "loss": 0.8019,
315
+ "step": 4500
316
+ },
317
+ {
318
+ "epoch": 15.24,
319
+ "learning_rate": 0.0001935701906412478,
320
+ "loss": 0.8,
321
+ "step": 4600
322
+ },
323
+ {
324
+ "epoch": 15.58,
325
+ "learning_rate": 0.00019097053726169842,
326
+ "loss": 0.7909,
327
+ "step": 4700
328
+ },
329
+ {
330
+ "epoch": 15.91,
331
+ "learning_rate": 0.00018837088388214902,
332
+ "loss": 0.7941,
333
+ "step": 4800
334
+ },
335
+ {
336
+ "epoch": 16.24,
337
+ "learning_rate": 0.0001857712305025996,
338
+ "loss": 0.7888,
339
+ "step": 4900
340
+ },
341
+ {
342
+ "epoch": 16.57,
343
+ "learning_rate": 0.00018317157712305026,
344
+ "loss": 0.767,
345
+ "step": 5000
346
+ },
347
+ {
348
+ "epoch": 16.57,
349
+ "eval_cer": 0.14326587352903142,
350
+ "eval_loss": 0.5884789228439331,
351
+ "eval_runtime": 143.5148,
352
+ "eval_samples_per_second": 33.718,
353
+ "eval_steps_per_second": 2.111,
354
+ "step": 5000
355
+ },
356
+ {
357
+ "epoch": 16.9,
358
+ "learning_rate": 0.00018057192374350086,
359
+ "loss": 0.7655,
360
+ "step": 5100
361
+ },
362
+ {
363
+ "epoch": 17.23,
364
+ "learning_rate": 0.00017797227036395145,
365
+ "loss": 0.7571,
366
+ "step": 5200
367
+ },
368
+ {
369
+ "epoch": 17.56,
370
+ "learning_rate": 0.00017537261698440207,
371
+ "loss": 0.754,
372
+ "step": 5300
373
+ },
374
+ {
375
+ "epoch": 17.9,
376
+ "learning_rate": 0.00017277296360485267,
377
+ "loss": 0.764,
378
+ "step": 5400
379
+ },
380
+ {
381
+ "epoch": 18.23,
382
+ "learning_rate": 0.0001701733102253033,
383
+ "loss": 0.7461,
384
+ "step": 5500
385
+ },
386
+ {
387
+ "epoch": 18.56,
388
+ "learning_rate": 0.00016757365684575388,
389
+ "loss": 0.7363,
390
+ "step": 5600
391
+ },
392
+ {
393
+ "epoch": 18.89,
394
+ "learning_rate": 0.00016497400346620448,
395
+ "loss": 0.7315,
396
+ "step": 5700
397
+ },
398
+ {
399
+ "epoch": 19.22,
400
+ "learning_rate": 0.0001623743500866551,
401
+ "loss": 0.7193,
402
+ "step": 5800
403
+ },
404
+ {
405
+ "epoch": 19.55,
406
+ "learning_rate": 0.00015977469670710572,
407
+ "loss": 0.7253,
408
+ "step": 5900
409
+ },
410
+ {
411
+ "epoch": 19.88,
412
+ "learning_rate": 0.00015717504332755632,
413
+ "loss": 0.7201,
414
+ "step": 6000
415
+ },
416
+ {
417
+ "epoch": 19.88,
418
+ "eval_cer": 0.1377699469804733,
419
+ "eval_loss": 0.5682498812675476,
420
+ "eval_runtime": 145.5022,
421
+ "eval_samples_per_second": 33.257,
422
+ "eval_steps_per_second": 2.082,
423
+ "step": 6000
424
+ },
425
+ {
426
+ "epoch": 20.22,
427
+ "learning_rate": 0.00015457538994800694,
428
+ "loss": 0.7203,
429
+ "step": 6100
430
+ },
431
+ {
432
+ "epoch": 20.55,
433
+ "learning_rate": 0.00015197573656845753,
434
+ "loss": 0.7045,
435
+ "step": 6200
436
+ },
437
+ {
438
+ "epoch": 20.88,
439
+ "learning_rate": 0.00014937608318890813,
440
+ "loss": 0.7072,
441
+ "step": 6300
442
+ },
443
+ {
444
+ "epoch": 21.21,
445
+ "learning_rate": 0.00014677642980935872,
446
+ "loss": 0.6979,
447
+ "step": 6400
448
+ },
449
+ {
450
+ "epoch": 21.54,
451
+ "learning_rate": 0.00014417677642980934,
452
+ "loss": 0.6971,
453
+ "step": 6500
454
+ },
455
+ {
456
+ "epoch": 21.87,
457
+ "learning_rate": 0.00014157712305025997,
458
+ "loss": 0.702,
459
+ "step": 6600
460
+ },
461
+ {
462
+ "epoch": 22.2,
463
+ "learning_rate": 0.00013897746967071056,
464
+ "loss": 0.6767,
465
+ "step": 6700
466
+ },
467
+ {
468
+ "epoch": 22.54,
469
+ "learning_rate": 0.00013637781629116116,
470
+ "loss": 0.682,
471
+ "step": 6800
472
+ },
473
+ {
474
+ "epoch": 22.87,
475
+ "learning_rate": 0.00013377816291161178,
476
+ "loss": 0.6729,
477
+ "step": 6900
478
+ },
479
+ {
480
+ "epoch": 23.2,
481
+ "learning_rate": 0.00013120450606585787,
482
+ "loss": 0.664,
483
+ "step": 7000
484
+ },
485
+ {
486
+ "epoch": 23.2,
487
+ "eval_cer": 0.13325197206776154,
488
+ "eval_loss": 0.5583386421203613,
489
+ "eval_runtime": 143.6376,
490
+ "eval_samples_per_second": 33.689,
491
+ "eval_steps_per_second": 2.109,
492
+ "step": 7000
493
+ },
494
+ {
495
+ "epoch": 23.53,
496
+ "learning_rate": 0.0001286048526863085,
497
+ "loss": 0.668,
498
+ "step": 7100
499
+ },
500
+ {
501
+ "epoch": 23.86,
502
+ "learning_rate": 0.00012600519930675908,
503
+ "loss": 0.67,
504
+ "step": 7200
505
+ },
506
+ {
507
+ "epoch": 24.19,
508
+ "learning_rate": 0.0001234055459272097,
509
+ "loss": 0.6608,
510
+ "step": 7300
511
+ },
512
+ {
513
+ "epoch": 24.52,
514
+ "learning_rate": 0.0001208058925476603,
515
+ "loss": 0.6543,
516
+ "step": 7400
517
+ },
518
+ {
519
+ "epoch": 24.86,
520
+ "learning_rate": 0.0001182062391681109,
521
+ "loss": 0.6648,
522
+ "step": 7500
523
+ },
524
+ {
525
+ "epoch": 25.19,
526
+ "learning_rate": 0.00011560658578856152,
527
+ "loss": 0.6473,
528
+ "step": 7600
529
+ },
530
+ {
531
+ "epoch": 25.52,
532
+ "learning_rate": 0.00011300693240901212,
533
+ "loss": 0.6546,
534
+ "step": 7700
535
+ },
536
+ {
537
+ "epoch": 25.85,
538
+ "learning_rate": 0.00011040727902946273,
539
+ "loss": 0.6397,
540
+ "step": 7800
541
+ },
542
+ {
543
+ "epoch": 26.18,
544
+ "learning_rate": 0.00010780762564991333,
545
+ "loss": 0.6395,
546
+ "step": 7900
547
+ },
548
+ {
549
+ "epoch": 26.51,
550
+ "learning_rate": 0.00010520797227036395,
551
+ "loss": 0.6296,
552
+ "step": 8000
553
+ },
554
+ {
555
+ "epoch": 26.51,
556
+ "eval_cer": 0.1298170179749127,
557
+ "eval_loss": 0.5415648818016052,
558
+ "eval_runtime": 142.9992,
559
+ "eval_samples_per_second": 33.839,
560
+ "eval_steps_per_second": 2.119,
561
+ "step": 8000
562
+ },
563
+ {
564
+ "epoch": 26.84,
565
+ "learning_rate": 0.00010260831889081456,
566
+ "loss": 0.6284,
567
+ "step": 8100
568
+ },
569
+ {
570
+ "epoch": 27.17,
571
+ "learning_rate": 0.00010000866551126515,
572
+ "loss": 0.6328,
573
+ "step": 8200
574
+ },
575
+ {
576
+ "epoch": 27.51,
577
+ "learning_rate": 9.740901213171576e-05,
578
+ "loss": 0.6233,
579
+ "step": 8300
580
+ },
581
+ {
582
+ "epoch": 27.84,
583
+ "learning_rate": 9.480935875216638e-05,
584
+ "loss": 0.6226,
585
+ "step": 8400
586
+ },
587
+ {
588
+ "epoch": 28.17,
589
+ "learning_rate": 9.220970537261698e-05,
590
+ "loss": 0.6214,
591
+ "step": 8500
592
+ },
593
+ {
594
+ "epoch": 28.5,
595
+ "learning_rate": 8.961005199306758e-05,
596
+ "loss": 0.6195,
597
+ "step": 8600
598
+ },
599
+ {
600
+ "epoch": 28.83,
601
+ "learning_rate": 8.701039861351819e-05,
602
+ "loss": 0.6138,
603
+ "step": 8700
604
+ },
605
+ {
606
+ "epoch": 29.16,
607
+ "learning_rate": 8.441074523396879e-05,
608
+ "loss": 0.6067,
609
+ "step": 8800
610
+ },
611
+ {
612
+ "epoch": 29.49,
613
+ "learning_rate": 8.181109185441941e-05,
614
+ "loss": 0.6076,
615
+ "step": 8900
616
+ },
617
+ {
618
+ "epoch": 29.83,
619
+ "learning_rate": 7.921143847487002e-05,
620
+ "loss": 0.6021,
621
+ "step": 9000
622
+ },
623
+ {
624
+ "epoch": 29.83,
625
+ "eval_cer": 0.12716604163972586,
626
+ "eval_loss": 0.5376999974250793,
627
+ "eval_runtime": 142.2332,
628
+ "eval_samples_per_second": 34.022,
629
+ "eval_steps_per_second": 2.13,
630
+ "step": 9000
631
+ },
632
+ {
633
+ "epoch": 30.16,
634
+ "learning_rate": 7.661178509532061e-05,
635
+ "loss": 0.5996,
636
+ "step": 9100
637
+ },
638
+ {
639
+ "epoch": 30.49,
640
+ "learning_rate": 7.401213171577122e-05,
641
+ "loss": 0.5989,
642
+ "step": 9200
643
+ },
644
+ {
645
+ "epoch": 30.82,
646
+ "learning_rate": 7.141247833622183e-05,
647
+ "loss": 0.5924,
648
+ "step": 9300
649
+ },
650
+ {
651
+ "epoch": 31.15,
652
+ "learning_rate": 6.881282495667244e-05,
653
+ "loss": 0.5858,
654
+ "step": 9400
655
+ },
656
+ {
657
+ "epoch": 31.48,
658
+ "learning_rate": 6.621317157712304e-05,
659
+ "loss": 0.5846,
660
+ "step": 9500
661
+ },
662
+ {
663
+ "epoch": 31.81,
664
+ "learning_rate": 6.363951473136915e-05,
665
+ "loss": 0.5906,
666
+ "step": 9600
667
+ },
668
+ {
669
+ "epoch": 32.15,
670
+ "learning_rate": 6.103986135181975e-05,
671
+ "loss": 0.5823,
672
+ "step": 9700
673
+ },
674
+ {
675
+ "epoch": 32.48,
676
+ "learning_rate": 5.8466204506065855e-05,
677
+ "loss": 0.5853,
678
+ "step": 9800
679
+ },
680
+ {
681
+ "epoch": 32.81,
682
+ "learning_rate": 5.586655112651646e-05,
683
+ "loss": 0.5759,
684
+ "step": 9900
685
+ },
686
+ {
687
+ "epoch": 33.14,
688
+ "learning_rate": 5.3266897746967065e-05,
689
+ "loss": 0.568,
690
+ "step": 10000
691
+ },
692
+ {
693
+ "epoch": 33.14,
694
+ "eval_cer": 0.1245797232639338,
695
+ "eval_loss": 0.5240569710731506,
696
+ "eval_runtime": 143.5422,
697
+ "eval_samples_per_second": 33.711,
698
+ "eval_steps_per_second": 2.111,
699
+ "step": 10000
700
+ },
701
+ {
702
+ "epoch": 33.47,
703
+ "learning_rate": 5.066724436741767e-05,
704
+ "loss": 0.574,
705
+ "step": 10100
706
+ },
707
+ {
708
+ "epoch": 33.8,
709
+ "learning_rate": 4.806759098786828e-05,
710
+ "loss": 0.5682,
711
+ "step": 10200
712
+ },
713
+ {
714
+ "epoch": 34.13,
715
+ "learning_rate": 4.546793760831888e-05,
716
+ "loss": 0.5636,
717
+ "step": 10300
718
+ },
719
+ {
720
+ "epoch": 34.47,
721
+ "learning_rate": 4.28682842287695e-05,
722
+ "loss": 0.5651,
723
+ "step": 10400
724
+ },
725
+ {
726
+ "epoch": 34.8,
727
+ "learning_rate": 4.02686308492201e-05,
728
+ "loss": 0.5693,
729
+ "step": 10500
730
+ },
731
+ {
732
+ "epoch": 35.13,
733
+ "learning_rate": 3.766897746967071e-05,
734
+ "loss": 0.5562,
735
+ "step": 10600
736
+ },
737
+ {
738
+ "epoch": 35.46,
739
+ "learning_rate": 3.5069324090121316e-05,
740
+ "loss": 0.5564,
741
+ "step": 10700
742
+ },
743
+ {
744
+ "epoch": 35.79,
745
+ "learning_rate": 3.2469670710571924e-05,
746
+ "loss": 0.5613,
747
+ "step": 10800
748
+ },
749
+ {
750
+ "epoch": 36.12,
751
+ "learning_rate": 2.9870017331022526e-05,
752
+ "loss": 0.5519,
753
+ "step": 10900
754
+ },
755
+ {
756
+ "epoch": 36.45,
757
+ "learning_rate": 2.7270363951473134e-05,
758
+ "loss": 0.5519,
759
+ "step": 11000
760
+ },
761
+ {
762
+ "epoch": 36.45,
763
+ "eval_cer": 0.12275313591103065,
764
+ "eval_loss": 0.5184260606765747,
765
+ "eval_runtime": 142.3486,
766
+ "eval_samples_per_second": 33.994,
767
+ "eval_steps_per_second": 2.129,
768
+ "step": 11000
769
+ },
770
+ {
771
+ "epoch": 36.79,
772
+ "learning_rate": 2.4670710571923742e-05,
773
+ "loss": 0.5605,
774
+ "step": 11100
775
+ },
776
+ {
777
+ "epoch": 37.12,
778
+ "learning_rate": 2.207105719237435e-05,
779
+ "loss": 0.5512,
780
+ "step": 11200
781
+ },
782
+ {
783
+ "epoch": 37.45,
784
+ "learning_rate": 1.9471403812824955e-05,
785
+ "loss": 0.5539,
786
+ "step": 11300
787
+ },
788
+ {
789
+ "epoch": 37.78,
790
+ "learning_rate": 1.687175043327556e-05,
791
+ "loss": 0.5467,
792
+ "step": 11400
793
+ },
794
+ {
795
+ "epoch": 38.11,
796
+ "learning_rate": 1.4272097053726168e-05,
797
+ "loss": 0.5464,
798
+ "step": 11500
799
+ },
800
+ {
801
+ "epoch": 38.44,
802
+ "learning_rate": 1.1672443674176775e-05,
803
+ "loss": 0.5457,
804
+ "step": 11600
805
+ },
806
+ {
807
+ "epoch": 38.77,
808
+ "learning_rate": 9.072790294627383e-06,
809
+ "loss": 0.5445,
810
+ "step": 11700
811
+ },
812
+ {
813
+ "epoch": 39.11,
814
+ "learning_rate": 6.473136915077989e-06,
815
+ "loss": 0.5395,
816
+ "step": 11800
817
+ },
818
+ {
819
+ "epoch": 39.44,
820
+ "learning_rate": 3.8734835355285955e-06,
821
+ "loss": 0.542,
822
+ "step": 11900
823
+ },
824
+ {
825
+ "epoch": 39.77,
826
+ "learning_rate": 1.2738301559792027e-06,
827
+ "loss": 0.5395,
828
+ "step": 12000
829
+ },
830
+ {
831
+ "epoch": 39.77,
832
+ "eval_cer": 0.12272080693133325,
833
+ "eval_loss": 0.5156292915344238,
834
+ "eval_runtime": 142.2326,
835
+ "eval_samples_per_second": 34.022,
836
+ "eval_steps_per_second": 2.13,
837
+ "step": 12000
838
+ }
839
+ ],
840
+ "logging_steps": 100,
841
+ "max_steps": 12040,
842
+ "num_train_epochs": 40,
843
+ "save_steps": 1000,
844
+ "total_flos": 1.9494106689613857e+20,
845
+ "trial_name": null,
846
+ "trial_params": null
847
+ }
checkpoint-12000/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:651a4f583ecd7f8a3a536ce2eb86832f1e9a2919291d8f9f5361918ab4c980fe
3
+ size 3963
checkpoint-12000/vocab.json ADDED
@@ -0,0 +1,1432 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "0": 1,
3
+ "1": 2,
4
+ "2": 3,
5
+ "3": 4,
6
+ "4": 5,
7
+ "5": 6,
8
+ "6": 7,
9
+ "7": 8,
10
+ "8": 9,
11
+ "9": 10,
12
+ "[PAD]": 1429,
13
+ "[UNK]": 1428,
14
+ "a": 11,
15
+ "b": 12,
16
+ "c": 13,
17
+ "d": 14,
18
+ "e": 15,
19
+ "f": 16,
20
+ "g": 17,
21
+ "h": 18,
22
+ "i": 19,
23
+ "k": 20,
24
+ "l": 21,
25
+ "m": 22,
26
+ "n": 23,
27
+ "o": 24,
28
+ "p": 25,
29
+ "r": 26,
30
+ "s": 27,
31
+ "t": 28,
32
+ "u": 29,
33
+ "v": 30,
34
+ "w": 31,
35
+ "x": 32,
36
+ "y": 33,
37
+ "z": 34,
38
+ "|": 0,
39
+ "ㄱ": 35,
40
+ "ㄴ": 36,
41
+ "ㄹ": 37,
42
+ "ㅁ": 38,
43
+ "ㅅ": 39,
44
+ "ㅇ": 40,
45
+ "ㅈ": 41,
46
+ "ㅡ": 42,
47
+ "가": 43,
48
+ "각": 44,
49
+ "간": 45,
50
+ "갇": 46,
51
+ "갈": 47,
52
+ "감": 48,
53
+ "갑": 49,
54
+ "값": 50,
55
+ "갓": 51,
56
+ "갔": 52,
57
+ "강": 53,
58
+ "갖": 54,
59
+ "같": 55,
60
+ "갚": 56,
61
+ "갛": 57,
62
+ "개": 58,
63
+ "객": 59,
64
+ "갠": 60,
65
+ "갤": 61,
66
+ "갭": 62,
67
+ "갯": 63,
68
+ "갰": 64,
69
+ "갱": 65,
70
+ "갸": 66,
71
+ "걀": 67,
72
+ "걍": 68,
73
+ "걔": 69,
74
+ "거": 70,
75
+ "걱": 71,
76
+ "건": 72,
77
+ "걷": 73,
78
+ "걸": 74,
79
+ "검": 75,
80
+ "겁": 76,
81
+ "것": 77,
82
+ "겉": 78,
83
+ "게": 79,
84
+ "겐": 80,
85
+ "겔": 81,
86
+ "겠": 82,
87
+ "겨": 83,
88
+ "격": 84,
89
+ "겪": 85,
90
+ "견": 86,
91
+ "결": 87,
92
+ "겸": 88,
93
+ "겹": 89,
94
+ "겼": 90,
95
+ "경": 91,
96
+ "곁": 92,
97
+ "계": 93,
98
+ "고": 94,
99
+ "곡": 95,
100
+ "곤": 96,
101
+ "곧": 97,
102
+ "골": 98,
103
+ "곰": 99,
104
+ "곱": 100,
105
+ "곳": 101,
106
+ "공": 102,
107
+ "과": 103,
108
+ "곽": 104,
109
+ "관": 105,
110
+ "괄": 106,
111
+ "괍": 107,
112
+ "광": 108,
113
+ "괜": 109,
114
+ "괴": 110,
115
+ "굉": 111,
116
+ "교": 112,
117
+ "굥": 113,
118
+ "구": 114,
119
+ "국": 115,
120
+ "군": 116,
121
+ "굳": 117,
122
+ "굴": 118,
123
+ "굵": 119,
124
+ "굶": 120,
125
+ "굽": 121,
126
+ "굿": 122,
127
+ "궁": 123,
128
+ "궈": 124,
129
+ "권": 125,
130
+ "궜": 126,
131
+ "궤": 127,
132
+ "귀": 128,
133
+ "규": 129,
134
+ "균": 130,
135
+ "귤": 131,
136
+ "그": 132,
137
+ "극": 133,
138
+ "근": 134,
139
+ "귿": 135,
140
+ "글": 136,
141
+ "긁": 137,
142
+ "금": 138,
143
+ "급": 139,
144
+ "긋": 140,
145
+ "긍": 141,
146
+ "긎": 142,
147
+ "기": 143,
148
+ "긴": 144,
149
+ "길": 145,
150
+ "김": 146,
151
+ "깁": 147,
152
+ "깃": 148,
153
+ "깊": 149,
154
+ "까": 150,
155
+ "깍": 151,
156
+ "깐": 152,
157
+ "깔": 153,
158
+ "깜": 154,
159
+ "깝": 155,
160
+ "깠": 156,
161
+ "깡": 157,
162
+ "깥": 158,
163
+ "깨": 159,
164
+ "깬": 160,
165
+ "깻": 161,
166
+ "깽": 162,
167
+ "꺄": 163,
168
+ "꺼": 164,
169
+ "꺽": 165,
170
+ "껄": 166,
171
+ "껌": 167,
172
+ "껍": 168,
173
+ "껏": 169,
174
+ "껐": 170,
175
+ "껑": 171,
176
+ "께": 172,
177
+ "껴": 173,
178
+ "꼈": 174,
179
+ "꼐": 175,
180
+ "꼬": 176,
181
+ "꼭": 177,
182
+ "꼰": 178,
183
+ "꼴": 179,
184
+ "꼼": 180,
185
+ "꼽": 181,
186
+ "꽂": 182,
187
+ "꽃": 183,
188
+ "꽈": 184,
189
+ "꽉": 185,
190
+ "꽤": 186,
191
+ "꾸": 187,
192
+ "꾹": 188,
193
+ "꾼": 189,
194
+ "꿀": 190,
195
+ "꿈": 191,
196
+ "꿍": 192,
197
+ "꿔": 193,
198
+ "꿨": 194,
199
+ "꿰": 195,
200
+ "뀌": 196,
201
+ "뀐": 197,
202
+ "뀔": 198,
203
+ "뀝": 199,
204
+ "끄": 200,
205
+ "끈": 201,
206
+ "끊": 202,
207
+ "끌": 203,
208
+ "끓": 204,
209
+ "끔": 205,
210
+ "끕": 206,
211
+ "끗": 207,
212
+ "끝": 208,
213
+ "끼": 209,
214
+ "끽": 210,
215
+ "낀": 211,
216
+ "낄": 212,
217
+ "낌": 213,
218
+ "낑": 214,
219
+ "나": 215,
220
+ "낙": 216,
221
+ "낚": 217,
222
+ "난": 218,
223
+ "날": 219,
224
+ "낡": 220,
225
+ "남": 221,
226
+ "납": 222,
227
+ "낫": 223,
228
+ "났": 224,
229
+ "낭": 225,
230
+ "낮": 226,
231
+ "낯": 227,
232
+ "낱": 228,
233
+ "낳": 229,
234
+ "내": 230,
235
+ "낸": 231,
236
+ "낼": 232,
237
+ "냄": 233,
238
+ "냅": 234,
239
+ "냈": 235,
240
+ "냉": 236,
241
+ "냐": 237,
242
+ "냠": 238,
243
+ "냥": 239,
244
+ "너": 240,
245
+ "넉": 241,
246
+ "넌": 242,
247
+ "널": 243,
248
+ "넓": 244,
249
+ "넘": 245,
250
+ "넛": 246,
251
+ "넣": 247,
252
+ "네": 248,
253
+ "넥": 249,
254
+ "넬": 250,
255
+ "넴": 251,
256
+ "넷": 252,
257
+ "녀": 253,
258
+ "녁": 254,
259
+ "년": 255,
260
+ "념": 256,
261
+ "녔": 257,
262
+ "녕": 258,
263
+ "노": 259,
264
+ "녹": 260,
265
+ "논": 261,
266
+ "놀": 262,
267
+ "놈": 263,
268
+ "농": 264,
269
+ "높": 265,
270
+ "놓": 266,
271
+ "놔": 267,
272
+ "놨": 268,
273
+ "뇌": 269,
274
+ "뇨": 270,
275
+ "누": 271,
276
+ "눅": 272,
277
+ "눈": 273,
278
+ "눌": 274,
279
+ "눔": 275,
280
+ "눕": 276,
281
+ "눗": 277,
282
+ "눠": 278,
283
+ "눴": 279,
284
+ "뉘": 280,
285
+ "뉜": 281,
286
+ "뉩": 282,
287
+ "뉴": 283,
288
+ "느": 284,
289
+ "는": 285,
290
+ "늘": 286,
291
+ "늙": 287,
292
+ "늠": 288,
293
+ "늡": 289,
294
+ "능": 290,
295
+ "늦": 291,
296
+ "늬": 292,
297
+ "니": 293,
298
+ "닉": 294,
299
+ "닌": 295,
300
+ "닐": 296,
301
+ "님": 297,
302
+ "닙": 298,
303
+ "닝": 299,
304
+ "다": 300,
305
+ "닥": 301,
306
+ "닦": 302,
307
+ "단": 303,
308
+ "닫": 304,
309
+ "달": 305,
310
+ "닭": 306,
311
+ "닮": 307,
312
+ "담": 308,
313
+ "답": 309,
314
+ "닷": 310,
315
+ "당": 311,
316
+ "닿": 312,
317
+ "대": 313,
318
+ "댁": 314,
319
+ "댄": 315,
320
+ "댈": 316,
321
+ "댐": 317,
322
+ "댑": 318,
323
+ "댓": 319,
324
+ "댔": 320,
325
+ "댕": 321,
326
+ "댜": 322,
327
+ "더": 323,
328
+ "덕": 324,
329
+ "던": 325,
330
+ "덜": 326,
331
+ "덟": 327,
332
+ "덤": 328,
333
+ "덥": 329,
334
+ "덧": 330,
335
+ "덩": 331,
336
+ "덮": 332,
337
+ "데": 333,
338
+ "덱": 334,
339
+ "델": 335,
340
+ "뎀": 336,
341
+ "뎅": 337,
342
+ "뎌": 338,
343
+ "도": 339,
344
+ "독": 340,
345
+ "돈": 341,
346
+ "돋": 342,
347
+ "돌": 343,
348
+ "돕": 344,
349
+ "돗": 345,
350
+ "동": 346,
351
+ "돼": 347,
352
+ "됐": 348,
353
+ "되": 349,
354
+ "된": 350,
355
+ "될": 351,
356
+ "됨": 352,
357
+ "됩": 353,
358
+ "됬": 354,
359
+ "두": 355,
360
+ "둑": 356,
361
+ "둔": 357,
362
+ "둘": 358,
363
+ "둠": 359,
364
+ "둡": 360,
365
+ "둥": 361,
366
+ "둬": 362,
367
+ "뒀": 363,
368
+ "뒤": 364,
369
+ "뒷": 365,
370
+ "듀": 366,
371
+ "드": 367,
372
+ "득": 368,
373
+ "든": 369,
374
+ "듣": 370,
375
+ "들": 371,
376
+ "듦": 372,
377
+ "듬": 373,
378
+ "듭": 374,
379
+ "듯": 375,
380
+ "등": 376,
381
+ "디": 377,
382
+ "딘": 378,
383
+ "딜": 379,
384
+ "딤": 380,
385
+ "딥": 381,
386
+ "딨": 382,
387
+ "딩": 383,
388
+ "딪": 384,
389
+ "따": 385,
390
+ "딱": 386,
391
+ "딴": 387,
392
+ "딸": 388,
393
+ "땀": 389,
394
+ "땅": 390,
395
+ "때": 391,
396
+ "땐": 392,
397
+ "땜": 393,
398
+ "땠": 394,
399
+ "땡": 395,
400
+ "떄": 396,
401
+ "떠": 397,
402
+ "떡": 398,
403
+ "떤": 399,
404
+ "떨": 400,
405
+ "떴": 401,
406
+ "떵": 402,
407
+ "떻": 403,
408
+ "떼": 404,
409
+ "뗀": 405,
410
+ "뗐": 406,
411
+ "또": 407,
412
+ "똑": 408,
413
+ "똠": 409,
414
+ "똥": 410,
415
+ "뚜": 411,
416
+ "뚝": 412,
417
+ "뚤": 413,
418
+ "뚫": 414,
419
+ "뚱": 415,
420
+ "뛰": 416,
421
+ "뛸": 417,
422
+ "뜨": 418,
423
+ "뜩": 419,
424
+ "뜬": 420,
425
+ "뜯": 421,
426
+ "뜰": 422,
427
+ "뜸": 423,
428
+ "뜹": 424,
429
+ "뜻": 425,
430
+ "띄": 426,
431
+ "띈": 427,
432
+ "띌": 428,
433
+ "띕": 429,
434
+ "띠": 430,
435
+ "띡": 431,
436
+ "띵": 432,
437
+ "라": 433,
438
+ "락": 434,
439
+ "란": 435,
440
+ "랄": 436,
441
+ "람": 437,
442
+ "랍": 438,
443
+ "랏": 439,
444
+ "랐": 440,
445
+ "랑": 441,
446
+ "랗": 442,
447
+ "래": 443,
448
+ "랙": 444,
449
+ "랜": 445,
450
+ "램": 446,
451
+ "랩": 447,
452
+ "랫": 448,
453
+ "랬": 449,
454
+ "략": 450,
455
+ "럈": 451,
456
+ "량": 452,
457
+ "러": 453,
458
+ "럭": 454,
459
+ "런": 455,
460
+ "럴": 456,
461
+ "럼": 457,
462
+ "럽": 458,
463
+ "렀": 459,
464
+ "렁": 460,
465
+ "렇": 461,
466
+ "레": 462,
467
+ "렉": 463,
468
+ "렌": 464,
469
+ "렘": 465,
470
+ "렛": 466,
471
+ "려": 467,
472
+ "력": 468,
473
+ "련": 469,
474
+ "렬": 470,
475
+ "렴": 471,
476
+ "렵": 472,
477
+ "렷": 473,
478
+ "렸": 474,
479
+ "령": 475,
480
+ "례": 476,
481
+ "로": 477,
482
+ "록": 478,
483
+ "론": 479,
484
+ "롤": 480,
485
+ "롬": 481,
486
+ "롭": 482,
487
+ "롯": 483,
488
+ "롱": 484,
489
+ "롷": 485,
490
+ "뢰": 486,
491
+ "료": 487,
492
+ "룡": 488,
493
+ "루": 489,
494
+ "룩": 490,
495
+ "룬": 491,
496
+ "룰": 492,
497
+ "룹": 493,
498
+ "뤄": 494,
499
+ "뤘": 495,
500
+ "류": 496,
501
+ "률": 497,
502
+ "륨": 498,
503
+ "륭": 499,
504
+ "르": 500,
505
+ "륵": 501,
506
+ "른": 502,
507
+ "를": 503,
508
+ "름": 504,
509
+ "릅": 505,
510
+ "릇": 506,
511
+ "릉": 507,
512
+ "릎": 508,
513
+ "리": 509,
514
+ "릭": 510,
515
+ "린": 511,
516
+ "릴": 512,
517
+ "림": 513,
518
+ "립": 514,
519
+ "릿": 515,
520
+ "링": 516,
521
+ "마": 517,
522
+ "막": 518,
523
+ "만": 519,
524
+ "많": 520,
525
+ "말": 521,
526
+ "맑": 522,
527
+ "맘": 523,
528
+ "맙": 524,
529
+ "맛": 525,
530
+ "망": 526,
531
+ "맞": 527,
532
+ "맡": 528,
533
+ "맣": 529,
534
+ "매": 530,
535
+ "맥": 531,
536
+ "맨": 532,
537
+ "맴": 533,
538
+ "맵": 534,
539
+ "맷": 535,
540
+ "맸": 536,
541
+ "맹": 537,
542
+ "맺": 538,
543
+ "머": 539,
544
+ "먹": 540,
545
+ "먼": 541,
546
+ "멀": 542,
547
+ "멈": 543,
548
+ "멋": 544,
549
+ "멍": 545,
550
+ "메": 546,
551
+ "멕": 547,
552
+ "멘": 548,
553
+ "멜": 549,
554
+ "멤": 550,
555
+ "멧": 551,
556
+ "며": 552,
557
+ "면": 553,
558
+ "멸": 554,
559
+ "명": 555,
560
+ "몇": 556,
561
+ "모": 557,
562
+ "목": 558,
563
+ "몫": 559,
564
+ "몬": 560,
565
+ "몰": 561,
566
+ "몸": 562,
567
+ "몹": 563,
568
+ "못": 564,
569
+ "몽": 565,
570
+ "묘": 566,
571
+ "무": 567,
572
+ "묵": 568,
573
+ "묶": 569,
574
+ "문": 570,
575
+ "묻": 571,
576
+ "물": 572,
577
+ "묽": 573,
578
+ "뭇": 574,
579
+ "뭉": 575,
580
+ "뭐": 576,
581
+ "뭔": 577,
582
+ "뭘": 578,
583
+ "뭡": 579,
584
+ "뭣": 580,
585
+ "뮤": 581,
586
+ "므": 582,
587
+ "믄": 583,
588
+ "믈": 584,
589
+ "미": 585,
590
+ "믹": 586,
591
+ "민": 587,
592
+ "믿": 588,
593
+ "밀": 589,
594
+ "밈": 590,
595
+ "밌": 591,
596
+ "밍": 592,
597
+ "및": 593,
598
+ "밑": 594,
599
+ "바": 595,
600
+ "박": 596,
601
+ "밖": 597,
602
+ "반": 598,
603
+ "받": 599,
604
+ "발": 600,
605
+ "밝": 601,
606
+ "밟": 602,
607
+ "밤": 603,
608
+ "밥": 604,
609
+ "방": 605,
610
+ "배": 606,
611
+ "백": 607,
612
+ "밴": 608,
613
+ "밸": 609,
614
+ "뱁": 610,
615
+ "뱃": 611,
616
+ "뱅": 612,
617
+ "뱉": 613,
618
+ "버": 614,
619
+ "벅": 615,
620
+ "번": 616,
621
+ "벌": 617,
622
+ "범": 618,
623
+ "법": 619,
624
+ "벗": 620,
625
+ "벙": 621,
626
+ "벚": 622,
627
+ "베": 623,
628
+ "벡": 624,
629
+ "벤": 625,
630
+ "벨": 626,
631
+ "벳": 627,
632
+ "벼": 628,
633
+ "벽": 629,
634
+ "변": 630,
635
+ "별": 631,
636
+ "볍": 632,
637
+ "병": 633,
638
+ "볕": 634,
639
+ "보": 635,
640
+ "복": 636,
641
+ "볶": 637,
642
+ "본": 638,
643
+ "볼": 639,
644
+ "봄": 640,
645
+ "봅": 641,
646
+ "봈": 642,
647
+ "봉": 643,
648
+ "봐": 644,
649
+ "봤": 645,
650
+ "봬": 646,
651
+ "뵀": 647,
652
+ "뵌": 648,
653
+ "뵐": 649,
654
+ "뵙": 650,
655
+ "부": 651,
656
+ "북": 652,
657
+ "분": 653,
658
+ "붇": 654,
659
+ "불": 655,
660
+ "붉": 656,
661
+ "붓": 657,
662
+ "붙": 658,
663
+ "뷔": 659,
664
+ "뷰": 660,
665
+ "브": 661,
666
+ "븐": 662,
667
+ "블": 663,
668
+ "비": 664,
669
+ "빅": 665,
670
+ "빈": 666,
671
+ "빌": 667,
672
+ "빔": 668,
673
+ "빗": 669,
674
+ "빙": 670,
675
+ "빛": 671,
676
+ "빠": 672,
677
+ "빡": 673,
678
+ "빤": 674,
679
+ "빨": 675,
680
+ "빴": 676,
681
+ "빵": 677,
682
+ "빼": 678,
683
+ "빽": 679,
684
+ "뺀": 680,
685
+ "뺄": 681,
686
+ "뺏": 682,
687
+ "뺐": 683,
688
+ "뺑": 684,
689
+ "뺨": 685,
690
+ "뺭": 686,
691
+ "뻐": 687,
692
+ "뻑": 688,
693
+ "뻔": 689,
694
+ "뻗": 690,
695
+ "뻘": 691,
696
+ "뻣": 692,
697
+ "뻤": 693,
698
+ "뻥": 694,
699
+ "뼈": 695,
700
+ "뽀": 696,
701
+ "뽑": 697,
702
+ "뽕": 698,
703
+ "뾰": 699,
704
+ "뿅": 700,
705
+ "뿌": 701,
706
+ "뿍": 702,
707
+ "뿐": 703,
708
+ "뿔": 704,
709
+ "쁘": 705,
710
+ "쁜": 706,
711
+ "쁠": 707,
712
+ "쁨": 708,
713
+ "쁩": 709,
714
+ "삐": 710,
715
+ "삔": 711,
716
+ "삥": 712,
717
+ "사": 713,
718
+ "삭": 714,
719
+ "산": 715,
720
+ "살": 716,
721
+ "삶": 717,
722
+ "삼": 718,
723
+ "삽": 719,
724
+ "삿": 720,
725
+ "샀": 721,
726
+ "상": 722,
727
+ "새": 723,
728
+ "색": 724,
729
+ "샌": 725,
730
+ "샘": 726,
731
+ "생": 727,
732
+ "샤": 728,
733
+ "샥": 729,
734
+ "샴": 730,
735
+ "샵": 731,
736
+ "서": 732,
737
+ "석": 733,
738
+ "섞": 734,
739
+ "선": 735,
740
+ "설": 736,
741
+ "섬": 737,
742
+ "섭": 738,
743
+ "섯": 739,
744
+ "섰": 740,
745
+ "성": 741,
746
+ "세": 742,
747
+ "섹": 743,
748
+ "센": 744,
749
+ "셀": 745,
750
+ "셈": 746,
751
+ "셉": 747,
752
+ "셋": 748,
753
+ "셌": 749,
754
+ "셔": 750,
755
+ "션": 751,
756
+ "셜": 752,
757
+ "셨": 753,
758
+ "셰": 754,
759
+ "소": 755,
760
+ "속": 756,
761
+ "손": 757,
762
+ "솔": 758,
763
+ "솜": 759,
764
+ "솟": 760,
765
+ "송": 761,
766
+ "솥": 762,
767
+ "쇄": 763,
768
+ "쇠": 764,
769
+ "쇼": 765,
770
+ "숀": 766,
771
+ "숍": 767,
772
+ "수": 768,
773
+ "숙": 769,
774
+ "순": 770,
775
+ "숟": 771,
776
+ "술": 772,
777
+ "숨": 773,
778
+ "숫": 774,
779
+ "숭": 775,
780
+ "숯": 776,
781
+ "숱": 777,
782
+ "숲": 778,
783
+ "숴": 779,
784
+ "쉐": 780,
785
+ "쉬": 781,
786
+ "쉴": 782,
787
+ "쉽": 783,
788
+ "슈": 784,
789
+ "슉": 785,
790
+ "슘": 786,
791
+ "슛": 787,
792
+ "슝": 788,
793
+ "스": 789,
794
+ "슥": 790,
795
+ "슨": 791,
796
+ "슬": 792,
797
+ "슴": 793,
798
+ "습": 794,
799
+ "슷": 795,
800
+ "승": 796,
801
+ "시": 797,
802
+ "식": 798,
803
+ "신": 799,
804
+ "싣": 800,
805
+ "실": 801,
806
+ "싫": 802,
807
+ "심": 803,
808
+ "십": 804,
809
+ "싱": 805,
810
+ "싴": 806,
811
+ "싶": 807,
812
+ "싸": 808,
813
+ "싹": 809,
814
+ "싼": 810,
815
+ "쌀": 811,
816
+ "쌈": 812,
817
+ "쌉": 813,
818
+ "쌍": 814,
819
+ "쌓": 815,
820
+ "쌤": 816,
821
+ "쌩": 817,
822
+ "쌰": 818,
823
+ "썜": 819,
824
+ "써": 820,
825
+ "썩": 821,
826
+ "썬": 822,
827
+ "썰": 823,
828
+ "썸": 824,
829
+ "썹": 825,
830
+ "썼": 826,
831
+ "썽": 827,
832
+ "쎄": 828,
833
+ "쎈": 829,
834
+ "쎼": 830,
835
+ "쏘": 831,
836
+ "쏙": 832,
837
+ "쏜": 833,
838
+ "쏟": 834,
839
+ "쏠": 835,
840
+ "쏴": 836,
841
+ "쑉": 837,
842
+ "쑤": 838,
843
+ "쑥": 839,
844
+ "쓰": 840,
845
+ "쓱": 841,
846
+ "쓴": 842,
847
+ "쓸": 843,
848
+ "씀": 844,
849
+ "씁": 845,
850
+ "씌": 846,
851
+ "씨": 847,
852
+ "씩": 848,
853
+ "씬": 849,
854
+ "씹": 850,
855
+ "씻": 851,
856
+ "아": 852,
857
+ "악": 853,
858
+ "안": 854,
859
+ "앉": 855,
860
+ "않": 856,
861
+ "알": 857,
862
+ "앓": 858,
863
+ "암": 859,
864
+ "압": 860,
865
+ "앗": 861,
866
+ "았": 862,
867
+ "앙": 863,
868
+ "앞": 864,
869
+ "애": 865,
870
+ "액": 866,
871
+ "앤": 867,
872
+ "앨": 868,
873
+ "앰": 869,
874
+ "앱": 870,
875
+ "앴": 871,
876
+ "앵": 872,
877
+ "야": 873,
878
+ "약": 874,
879
+ "얀": 875,
880
+ "얄": 876,
881
+ "얇": 877,
882
+ "얌": 878,
883
+ "양": 879,
884
+ "얕": 880,
885
+ "얗": 881,
886
+ "얘": 882,
887
+ "얜": 883,
888
+ "어": 884,
889
+ "억": 885,
890
+ "언": 886,
891
+ "얹": 887,
892
+ "얻": 888,
893
+ "얼": 889,
894
+ "엄": 890,
895
+ "업": 891,
896
+ "없": 892,
897
+ "엇": 893,
898
+ "었": 894,
899
+ "엉": 895,
900
+ "엎": 896,
901
+ "에": 897,
902
+ "엑": 898,
903
+ "엔": 899,
904
+ "엘": 900,
905
+ "엠": 901,
906
+ "엥": 902,
907
+ "여": 903,
908
+ "역": 904,
909
+ "연": 905,
910
+ "열": 906,
911
+ "염": 907,
912
+ "엽": 908,
913
+ "엿": 909,
914
+ "였": 910,
915
+ "영": 911,
916
+ "옆": 912,
917
+ "옇": 913,
918
+ "예": 914,
919
+ "옛": 915,
920
+ "오": 916,
921
+ "옥": 917,
922
+ "온": 918,
923
+ "올": 919,
924
+ "옮": 920,
925
+ "옳": 921,
926
+ "옴": 922,
927
+ "옵": 923,
928
+ "옷": 924,
929
+ "옹": 925,
930
+ "와": 926,
931
+ "완": 927,
932
+ "왈": 928,
933
+ "왓": 929,
934
+ "왔": 930,
935
+ "왕": 931,
936
+ "왜": 932,
937
+ "왠": 933,
938
+ "외": 934,
939
+ "왼": 935,
940
+ "요": 936,
941
+ "욕": 937,
942
+ "용": 938,
943
+ "우": 939,
944
+ "욱": 940,
945
+ "운": 941,
946
+ "울": 942,
947
+ "움": 943,
948
+ "웁": 944,
949
+ "웃": 945,
950
+ "웅": 946,
951
+ "워": 947,
952
+ "원": 948,
953
+ "월": 949,
954
+ "웠": 950,
955
+ "웨": 951,
956
+ "웩": 952,
957
+ "웬": 953,
958
+ "웰": 954,
959
+ "웹": 955,
960
+ "위": 956,
961
+ "윈": 957,
962
+ "윗": 958,
963
+ "유": 959,
964
+ "육": 960,
965
+ "윤": 961,
966
+ "율": 962,
967
+ "융": 963,
968
+ "으": 964,
969
+ "윽": 965,
970
+ "은": 966,
971
+ "읃": 967,
972
+ "을": 968,
973
+ "음": 969,
974
+ "읍": 970,
975
+ "읏": 971,
976
+ "응": 972,
977
+ "의": 973,
978
+ "이": 974,
979
+ "익": 975,
980
+ "인": 976,
981
+ "일": 977,
982
+ "읽": 978,
983
+ "잃": 979,
984
+ "임": 980,
985
+ "입": 981,
986
+ "잇": 982,
987
+ "있": 983,
988
+ "잉": 984,
989
+ "잊": 985,
990
+ "잎": 986,
991
+ "자": 987,
992
+ "작": 988,
993
+ "잔": 989,
994
+ "잖": 990,
995
+ "잘": 991,
996
+ "잠": 992,
997
+ "잡": 993,
998
+ "잣": 994,
999
+ "잤": 995,
1000
+ "장": 996,
1001
+ "잦": 997,
1002
+ "재": 998,
1003
+ "잰": 999,
1004
+ "잼": 1000,
1005
+ "잽": 1001,
1006
+ "쟁": 1002,
1007
+ "쟈": 1003,
1008
+ "쟤": 1004,
1009
+ "저": 1005,
1010
+ "적": 1006,
1011
+ "전": 1007,
1012
+ "절": 1008,
1013
+ "젊": 1009,
1014
+ "점": 1010,
1015
+ "접": 1011,
1016
+ "젓": 1012,
1017
+ "정": 1013,
1018
+ "젖": 1014,
1019
+ "제": 1015,
1020
+ "젝": 1016,
1021
+ "젠": 1017,
1022
+ "젤": 1018,
1023
+ "젯": 1019,
1024
+ "져": 1020,
1025
+ "졌": 1021,
1026
+ "조": 1022,
1027
+ "족": 1023,
1028
+ "존": 1024,
1029
+ "졸": 1025,
1030
+ "좀": 1026,
1031
+ "좁": 1027,
1032
+ "종": 1028,
1033
+ "좋": 1029,
1034
+ "좌": 1030,
1035
+ "좟": 1031,
1036
+ "죄": 1032,
1037
+ "죠": 1033,
1038
+ "죵": 1034,
1039
+ "주": 1035,
1040
+ "죽": 1036,
1041
+ "준": 1037,
1042
+ "줄": 1038,
1043
+ "줌": 1039,
1044
+ "줍": 1040,
1045
+ "중": 1041,
1046
+ "줘": 1042,
1047
+ "줬": 1043,
1048
+ "쥐": 1044,
1049
+ "쥬": 1045,
1050
+ "즈": 1046,
1051
+ "즉": 1047,
1052
+ "즌": 1048,
1053
+ "즐": 1049,
1054
+ "즘": 1050,
1055
+ "증": 1051,
1056
+ "지": 1052,
1057
+ "직": 1053,
1058
+ "진": 1054,
1059
+ "질": 1055,
1060
+ "짐": 1056,
1061
+ "집": 1057,
1062
+ "짓": 1058,
1063
+ "징": 1059,
1064
+ "짖": 1060,
1065
+ "짙": 1061,
1066
+ "짚": 1062,
1067
+ "짜": 1063,
1068
+ "짝": 1064,
1069
+ "짠": 1065,
1070
+ "짤": 1066,
1071
+ "짧": 1067,
1072
+ "짬": 1068,
1073
+ "짭": 1069,
1074
+ "짰": 1070,
1075
+ "짱": 1071,
1076
+ "째": 1072,
1077
+ "쨋": 1073,
1078
+ "쨌": 1074,
1079
+ "쨍": 1075,
1080
+ "쨰": 1076,
1081
+ "쩌": 1077,
1082
+ "쩍": 1078,
1083
+ "쩐": 1079,
1084
+ "쩔": 1080,
1085
+ "쩜": 1081,
1086
+ "쩝": 1082,
1087
+ "쩠": 1083,
1088
+ "쩡": 1084,
1089
+ "쪼": 1085,
1090
+ "쪽": 1086,
1091
+ "쫀": 1087,
1092
+ "쫄": 1088,
1093
+ "쫌": 1089,
1094
+ "쫓": 1090,
1095
+ "쫙": 1091,
1096
+ "쫜": 1092,
1097
+ "쬐": 1093,
1098
+ "쭈": 1094,
1099
+ "쭉": 1095,
1100
+ "쭤": 1096,
1101
+ "쮸": 1097,
1102
+ "쯔": 1098,
1103
+ "쯤": 1099,
1104
+ "찌": 1100,
1105
+ "찍": 1101,
1106
+ "찐": 1102,
1107
+ "찔": 1103,
1108
+ "찜": 1104,
1109
+ "찝": 1105,
1110
+ "찡": 1106,
1111
+ "찢": 1107,
1112
+ "차": 1108,
1113
+ "착": 1109,
1114
+ "찬": 1110,
1115
+ "찮": 1111,
1116
+ "찰": 1112,
1117
+ "참": 1113,
1118
+ "찻": 1114,
1119
+ "찼": 1115,
1120
+ "창": 1116,
1121
+ "찾": 1117,
1122
+ "채": 1118,
1123
+ "책": 1119,
1124
+ "챌": 1120,
1125
+ "챕": 1121,
1126
+ "챙": 1122,
1127
+ "처": 1123,
1128
+ "척": 1124,
1129
+ "천": 1125,
1130
+ "철": 1126,
1131
+ "첨": 1127,
1132
+ "첩": 1128,
1133
+ "첫": 1129,
1134
+ "청": 1130,
1135
+ "체": 1131,
1136
+ "첼": 1132,
1137
+ "쳅": 1133,
1138
+ "쳐": 1134,
1139
+ "쳤": 1135,
1140
+ "쳬": 1136,
1141
+ "초": 1137,
1142
+ "촉": 1138,
1143
+ "촌": 1139,
1144
+ "촘": 1140,
1145
+ "촛": 1141,
1146
+ "총": 1142,
1147
+ "촤": 1143,
1148
+ "촥": 1144,
1149
+ "촬": 1145,
1150
+ "최": 1146,
1151
+ "쵸": 1147,
1152
+ "추": 1148,
1153
+ "축": 1149,
1154
+ "춘": 1150,
1155
+ "출": 1151,
1156
+ "춤": 1152,
1157
+ "춥": 1153,
1158
+ "춧": 1154,
1159
+ "충": 1155,
1160
+ "춰": 1156,
1161
+ "췄": 1157,
1162
+ "췌": 1158,
1163
+ "취": 1159,
1164
+ "츄": 1160,
1165
+ "츠": 1161,
1166
+ "측": 1162,
1167
+ "츨": 1163,
1168
+ "층": 1164,
1169
+ "치": 1165,
1170
+ "칙": 1166,
1171
+ "친": 1167,
1172
+ "칠": 1168,
1173
+ "침": 1169,
1174
+ "칩": 1170,
1175
+ "칫": 1171,
1176
+ "칭": 1172,
1177
+ "카": 1173,
1178
+ "칵": 1174,
1179
+ "칸": 1175,
1180
+ "칼": 1176,
1181
+ "캐": 1177,
1182
+ "캔": 1178,
1183
+ "캘": 1179,
1184
+ "캠": 1180,
1185
+ "캡": 1181,
1186
+ "캣": 1182,
1187
+ "캬": 1183,
1188
+ "커": 1184,
1189
+ "컥": 1185,
1190
+ "컨": 1186,
1191
+ "컬": 1187,
1192
+ "컴": 1188,
1193
+ "컵": 1189,
1194
+ "컷": 1190,
1195
+ "컸": 1191,
1196
+ "케": 1192,
1197
+ "켄": 1193,
1198
+ "켓": 1194,
1199
+ "켜": 1195,
1200
+ "켠": 1196,
1201
+ "켰": 1197,
1202
+ "코": 1198,
1203
+ "콕": 1199,
1204
+ "콘": 1200,
1205
+ "콜": 1201,
1206
+ "콤": 1202,
1207
+ "콥": 1203,
1208
+ "콧": 1204,
1209
+ "콩": 1205,
1210
+ "콸": 1206,
1211
+ "쾌": 1207,
1212
+ "쿠": 1208,
1213
+ "쿡": 1209,
1214
+ "쿤": 1210,
1215
+ "쿨": 1211,
1216
+ "쿵": 1212,
1217
+ "퀄": 1213,
1218
+ "퀴": 1214,
1219
+ "큐": 1215,
1220
+ "큘": 1216,
1221
+ "크": 1217,
1222
+ "큰": 1218,
1223
+ "클": 1219,
1224
+ "큼": 1220,
1225
+ "큽": 1221,
1226
+ "키": 1222,
1227
+ "킨": 1223,
1228
+ "킬": 1224,
1229
+ "킴": 1225,
1230
+ "킵": 1226,
1231
+ "킹": 1227,
1232
+ "타": 1228,
1233
+ "탁": 1229,
1234
+ "탄": 1230,
1235
+ "탈": 1231,
1236
+ "탐": 1232,
1237
+ "탑": 1233,
1238
+ "탓": 1234,
1239
+ "탔": 1235,
1240
+ "탕": 1236,
1241
+ "태": 1237,
1242
+ "택": 1238,
1243
+ "탠": 1239,
1244
+ "탱": 1240,
1245
+ "터": 1241,
1246
+ "턱": 1242,
1247
+ "턴": 1243,
1248
+ "털": 1244,
1249
+ "텁": 1245,
1250
+ "텅": 1246,
1251
+ "테": 1247,
1252
+ "텍": 1248,
1253
+ "텐": 1249,
1254
+ "텔": 1250,
1255
+ "템": 1251,
1256
+ "텝": 1252,
1257
+ "텼": 1253,
1258
+ "토": 1254,
1259
+ "톡": 1255,
1260
+ "톤": 1256,
1261
+ "톨": 1257,
1262
+ "톱": 1258,
1263
+ "통": 1259,
1264
+ "퇴": 1260,
1265
+ "투": 1261,
1266
+ "툭": 1262,
1267
+ "툴": 1263,
1268
+ "툼": 1264,
1269
+ "퉁": 1265,
1270
+ "튀": 1266,
1271
+ "튜": 1267,
1272
+ "트": 1268,
1273
+ "특": 1269,
1274
+ "튼": 1270,
1275
+ "틀": 1271,
1276
+ "틈": 1272,
1277
+ "틍": 1273,
1278
+ "티": 1274,
1279
+ "틱": 1275,
1280
+ "틴": 1276,
1281
+ "틸": 1277,
1282
+ "팀": 1278,
1283
+ "팁": 1279,
1284
+ "팅": 1280,
1285
+ "파": 1281,
1286
+ "팍": 1282,
1287
+ "팎": 1283,
1288
+ "판": 1284,
1289
+ "팔": 1285,
1290
+ "팠": 1286,
1291
+ "팡": 1287,
1292
+ "팥": 1288,
1293
+ "패": 1289,
1294
+ "팩": 1290,
1295
+ "팬": 1291,
1296
+ "팽": 1292,
1297
+ "퍼": 1293,
1298
+ "퍽": 1294,
1299
+ "펀": 1295,
1300
+ "펄": 1296,
1301
+ "펌": 1297,
1302
+ "페": 1298,
1303
+ "펙": 1299,
1304
+ "펜": 1300,
1305
+ "펠": 1301,
1306
+ "펩": 1302,
1307
+ "펴": 1303,
1308
+ "편": 1304,
1309
+ "펼": 1305,
1310
+ "폈": 1306,
1311
+ "평": 1307,
1312
+ "폐": 1308,
1313
+ "포": 1309,
1314
+ "폭": 1310,
1315
+ "폰": 1311,
1316
+ "폴": 1312,
1317
+ "폼": 1313,
1318
+ "퐁": 1314,
1319
+ "표": 1315,
1320
+ "푸": 1316,
1321
+ "푹": 1317,
1322
+ "푼": 1318,
1323
+ "풀": 1319,
1324
+ "품": 1320,
1325
+ "풋": 1321,
1326
+ "풍": 1322,
1327
+ "퓨": 1323,
1328
+ "프": 1324,
1329
+ "픈": 1325,
1330
+ "플": 1326,
1331
+ "픔": 1327,
1332
+ "피": 1328,
1333
+ "픽": 1329,
1334
+ "핀": 1330,
1335
+ "필": 1331,
1336
+ "핍": 1332,
1337
+ "핏": 1333,
1338
+ "핑": 1334,
1339
+ "하": 1335,
1340
+ "학": 1336,
1341
+ "한": 1337,
1342
+ "할": 1338,
1343
+ "함": 1339,
1344
+ "합": 1340,
1345
+ "핫": 1341,
1346
+ "항": 1342,
1347
+ "해": 1343,
1348
+ "핵": 1344,
1349
+ "핸": 1345,
1350
+ "핼": 1346,
1351
+ "햄": 1347,
1352
+ "햅": 1348,
1353
+ "햇": 1349,
1354
+ "했": 1350,
1355
+ "행": 1351,
1356
+ "햍": 1352,
1357
+ "햐": 1353,
1358
+ "향": 1354,
1359
+ "헀": 1355,
1360
+ "허": 1356,
1361
+ "헉": 1357,
1362
+ "헌": 1358,
1363
+ "헐": 1359,
1364
+ "험": 1360,
1365
+ "헛": 1361,
1366
+ "헝": 1362,
1367
+ "헤": 1363,
1368
+ "헨": 1364,
1369
+ "헬": 1365,
1370
+ "헷": 1366,
1371
+ "헹": 1367,
1372
+ "혀": 1368,
1373
+ "혁": 1369,
1374
+ "현": 1370,
1375
+ "혈": 1371,
1376
+ "혐": 1372,
1377
+ "협": 1373,
1378
+ "혓": 1374,
1379
+ "혔": 1375,
1380
+ "형": 1376,
1381
+ "혜": 1377,
1382
+ "호": 1378,
1383
+ "혹": 1379,
1384
+ "혼": 1380,
1385
+ "홀": 1381,
1386
+ "홈": 1382,
1387
+ "홉": 1383,
1388
+ "홍": 1384,
1389
+ "화": 1385,
1390
+ "확": 1386,
1391
+ "환": 1387,
1392
+ "활": 1388,
1393
+ "홥": 1389,
1394
+ "황": 1390,
1395
+ "회": 1391,
1396
+ "획": 1392,
1397
+ "횟": 1393,
1398
+ "횡": 1394,
1399
+ "효": 1395,
1400
+ "횩": 1396,
1401
+ "후": 1397,
1402
+ "훅": 1398,
1403
+ "훈": 1399,
1404
+ "훌": 1400,
1405
+ "훑": 1401,
1406
+ "훔": 1402,
1407
+ "훨": 1403,
1408
+ "훼": 1404,
1409
+ "휑": 1405,
1410
+ "휘": 1406,
1411
+ "휙": 1407,
1412
+ "휩": 1408,
1413
+ "휴": 1409,
1414
+ "흉": 1410,
1415
+ "흐": 1411,
1416
+ "흑": 1412,
1417
+ "흔": 1413,
1418
+ "흘": 1414,
1419
+ "흙": 1415,
1420
+ "흠": 1416,
1421
+ "흡": 1417,
1422
+ "흣": 1418,
1423
+ "흥": 1419,
1424
+ "흩": 1420,
1425
+ "희": 1421,
1426
+ "흰": 1422,
1427
+ "히": 1423,
1428
+ "힌": 1424,
1429
+ "힐": 1425,
1430
+ "힘": 1426,
1431
+ "힙": 1427
1432
+ }
config.json ADDED
@@ -0,0 +1,117 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
3
+ "activation_dropout": 0.0,
4
+ "adapter_attn_dim": null,
5
+ "adapter_kernel_size": 3,
6
+ "adapter_stride": 2,
7
+ "add_adapter": false,
8
+ "apply_spec_augment": true,
9
+ "architectures": [
10
+ "Wav2Vec2ForCTC"
11
+ ],
12
+ "attention_dropout": 0.1,
13
+ "bos_token_id": 1,
14
+ "classifier_proj_size": 256,
15
+ "codevector_dim": 768,
16
+ "contrastive_logits_temperature": 0.1,
17
+ "conv_bias": true,
18
+ "conv_dim": [
19
+ 512,
20
+ 512,
21
+ 512,
22
+ 512,
23
+ 512,
24
+ 512,
25
+ 512
26
+ ],
27
+ "conv_kernel": [
28
+ 10,
29
+ 3,
30
+ 3,
31
+ 3,
32
+ 3,
33
+ 2,
34
+ 2
35
+ ],
36
+ "conv_stride": [
37
+ 5,
38
+ 2,
39
+ 2,
40
+ 2,
41
+ 2,
42
+ 2,
43
+ 2
44
+ ],
45
+ "ctc_loss_reduction": "mean",
46
+ "ctc_zero_infinity": false,
47
+ "diversity_loss_weight": 0.1,
48
+ "do_stable_layer_norm": true,
49
+ "eos_token_id": 2,
50
+ "feat_extract_activation": "gelu",
51
+ "feat_extract_dropout": 0.0,
52
+ "feat_extract_norm": "layer",
53
+ "feat_proj_dropout": 0.1,
54
+ "feat_quantizer_dropout": 0.0,
55
+ "final_dropout": 0.0,
56
+ "gradient_checkpointing": false,
57
+ "hidden_act": "gelu",
58
+ "hidden_dropout": 0.1,
59
+ "hidden_size": 1024,
60
+ "initializer_range": 0.02,
61
+ "intermediate_size": 4096,
62
+ "layer_norm_eps": 1e-05,
63
+ "layerdrop": 0.1,
64
+ "mask_channel_length": 10,
65
+ "mask_channel_min_space": 1,
66
+ "mask_channel_other": 0.0,
67
+ "mask_channel_prob": 0.0,
68
+ "mask_channel_selection": "static",
69
+ "mask_feature_length": 64,
70
+ "mask_feature_min_masks": 0,
71
+ "mask_feature_prob": 0.0,
72
+ "mask_time_length": 10,
73
+ "mask_time_min_masks": 2,
74
+ "mask_time_min_space": 1,
75
+ "mask_time_other": 0.0,
76
+ "mask_time_prob": 0.3,
77
+ "mask_time_selection": "static",
78
+ "model_type": "wav2vec2",
79
+ "num_adapter_layers": 3,
80
+ "num_attention_heads": 16,
81
+ "num_codevector_groups": 2,
82
+ "num_codevectors_per_group": 320,
83
+ "num_conv_pos_embedding_groups": 16,
84
+ "num_conv_pos_embeddings": 128,
85
+ "num_feat_extract_layers": 7,
86
+ "num_hidden_layers": 24,
87
+ "num_negatives": 100,
88
+ "output_hidden_size": 1024,
89
+ "pad_token_id": 1429,
90
+ "proj_codevector_dim": 768,
91
+ "tdnn_dilation": [
92
+ 1,
93
+ 2,
94
+ 3,
95
+ 1,
96
+ 1
97
+ ],
98
+ "tdnn_dim": [
99
+ 512,
100
+ 512,
101
+ 512,
102
+ 512,
103
+ 1500
104
+ ],
105
+ "tdnn_kernel": [
106
+ 5,
107
+ 3,
108
+ 3,
109
+ 1,
110
+ 1
111
+ ],
112
+ "torch_dtype": "float32",
113
+ "transformers_version": "4.33.0",
114
+ "use_weighted_layer_sum": false,
115
+ "vocab_size": 1432,
116
+ "xvector_output_dim": 512
117
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 39.9,
3
+ "eval_cer": 0.12280162938057675,
4
+ "eval_loss": 0.51564621925354,
5
+ "eval_runtime": 141.8188,
6
+ "eval_samples": 4839,
7
+ "eval_samples_per_second": 34.121,
8
+ "eval_steps_per_second": 2.137
9
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "feature_extractor_type": "Wav2Vec2FeatureExtractor",
4
+ "feature_size": 1,
5
+ "padding_side": "right",
6
+ "padding_value": 0,
7
+ "processor_class": "Wav2Vec2Processor",
8
+ "return_attention_mask": true,
9
+ "sampling_rate": 16000
10
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c6d305748d7ab8bad10dbad59e08e5a4133c1e8fbb880d0ee71623dde0db797a
3
+ size 1267773037
special_tokens_map.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": true,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ {
11
+ "content": "</s>",
12
+ "lstrip": false,
13
+ "normalized": true,
14
+ "rstrip": false,
15
+ "single_word": false
16
+ },
17
+ {
18
+ "content": "<s>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ {
25
+ "content": "</s>",
26
+ "lstrip": false,
27
+ "normalized": true,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ ],
32
+ "bos_token": "<s>",
33
+ "eos_token": "</s>",
34
+ "pad_token": "[PAD]",
35
+ "unk_token": "[UNK]"
36
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "clean_up_tokenization_spaces": true,
4
+ "config": null,
5
+ "do_lower_case": false,
6
+ "eos_token": "</s>",
7
+ "model_max_length": 1000000000000000019884624838656,
8
+ "pad_token": "[PAD]",
9
+ "processor_class": "Wav2Vec2Processor",
10
+ "replace_word_delimiter_char": " ",
11
+ "target_lang": null,
12
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
13
+ "tokenizer_type": "wav2vec2",
14
+ "trust_remote_code": false,
15
+ "unk_token": "[UNK]",
16
+ "word_delimiter_token": "|"
17
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 39.9,
3
+ "train_loss": 1.365207207163307,
4
+ "train_runtime": 27512.8703,
5
+ "train_samples": 38612,
6
+ "train_samples_per_second": 56.137,
7
+ "train_steps_per_second": 0.438
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,856 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 39.90057995028997,
5
+ "eval_steps": 1000,
6
+ "global_step": 12040,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.33,
13
+ "learning_rate": 5.6399999999999995e-05,
14
+ "loss": 37.3417,
15
+ "step": 100
16
+ },
17
+ {
18
+ "epoch": 0.66,
19
+ "learning_rate": 0.0001164,
20
+ "loss": 12.5064,
21
+ "step": 200
22
+ },
23
+ {
24
+ "epoch": 0.99,
25
+ "learning_rate": 0.00017639999999999998,
26
+ "loss": 4.823,
27
+ "step": 300
28
+ },
29
+ {
30
+ "epoch": 1.33,
31
+ "learning_rate": 0.0002364,
32
+ "loss": 4.7159,
33
+ "step": 400
34
+ },
35
+ {
36
+ "epoch": 1.66,
37
+ "learning_rate": 0.0002964,
38
+ "loss": 4.6828,
39
+ "step": 500
40
+ },
41
+ {
42
+ "epoch": 1.99,
43
+ "learning_rate": 0.0002975563258232235,
44
+ "loss": 4.5391,
45
+ "step": 600
46
+ },
47
+ {
48
+ "epoch": 2.32,
49
+ "learning_rate": 0.0002949566724436741,
50
+ "loss": 4.3365,
51
+ "step": 700
52
+ },
53
+ {
54
+ "epoch": 2.65,
55
+ "learning_rate": 0.0002923570190641247,
56
+ "loss": 3.2316,
57
+ "step": 800
58
+ },
59
+ {
60
+ "epoch": 2.98,
61
+ "learning_rate": 0.00028975736568457536,
62
+ "loss": 2.1607,
63
+ "step": 900
64
+ },
65
+ {
66
+ "epoch": 3.31,
67
+ "learning_rate": 0.00028715771230502596,
68
+ "loss": 1.778,
69
+ "step": 1000
70
+ },
71
+ {
72
+ "epoch": 3.31,
73
+ "eval_cer": 0.3050481701797491,
74
+ "eval_loss": 1.2772819995880127,
75
+ "eval_runtime": 143.2917,
76
+ "eval_samples_per_second": 33.77,
77
+ "eval_steps_per_second": 2.115,
78
+ "step": 1000
79
+ },
80
+ {
81
+ "epoch": 3.65,
82
+ "learning_rate": 0.00028455805892547655,
83
+ "loss": 1.6287,
84
+ "step": 1100
85
+ },
86
+ {
87
+ "epoch": 3.98,
88
+ "learning_rate": 0.0002819584055459272,
89
+ "loss": 1.5046,
90
+ "step": 1200
91
+ },
92
+ {
93
+ "epoch": 4.31,
94
+ "learning_rate": 0.0002793587521663778,
95
+ "loss": 1.4046,
96
+ "step": 1300
97
+ },
98
+ {
99
+ "epoch": 4.64,
100
+ "learning_rate": 0.0002767590987868284,
101
+ "loss": 1.3378,
102
+ "step": 1400
103
+ },
104
+ {
105
+ "epoch": 4.97,
106
+ "learning_rate": 0.000274159445407279,
107
+ "loss": 1.2868,
108
+ "step": 1500
109
+ },
110
+ {
111
+ "epoch": 5.3,
112
+ "learning_rate": 0.0002715597920277296,
113
+ "loss": 1.2376,
114
+ "step": 1600
115
+ },
116
+ {
117
+ "epoch": 5.63,
118
+ "learning_rate": 0.00026896013864818023,
119
+ "loss": 1.2126,
120
+ "step": 1700
121
+ },
122
+ {
123
+ "epoch": 5.97,
124
+ "learning_rate": 0.0002663604852686308,
125
+ "loss": 1.1781,
126
+ "step": 1800
127
+ },
128
+ {
129
+ "epoch": 6.3,
130
+ "learning_rate": 0.0002637608318890814,
131
+ "loss": 1.1426,
132
+ "step": 1900
133
+ },
134
+ {
135
+ "epoch": 6.63,
136
+ "learning_rate": 0.00026116117850953207,
137
+ "loss": 1.1037,
138
+ "step": 2000
139
+ },
140
+ {
141
+ "epoch": 6.63,
142
+ "eval_cer": 0.18878507694297167,
143
+ "eval_loss": 0.7715919017791748,
144
+ "eval_runtime": 143.3347,
145
+ "eval_samples_per_second": 33.76,
146
+ "eval_steps_per_second": 2.114,
147
+ "step": 2000
148
+ },
149
+ {
150
+ "epoch": 6.96,
151
+ "learning_rate": 0.00025856152512998266,
152
+ "loss": 1.0971,
153
+ "step": 2100
154
+ },
155
+ {
156
+ "epoch": 7.29,
157
+ "learning_rate": 0.00025596187175043326,
158
+ "loss": 1.0693,
159
+ "step": 2200
160
+ },
161
+ {
162
+ "epoch": 7.62,
163
+ "learning_rate": 0.00025336221837088385,
164
+ "loss": 1.0398,
165
+ "step": 2300
166
+ },
167
+ {
168
+ "epoch": 7.95,
169
+ "learning_rate": 0.00025076256499133445,
170
+ "loss": 1.0282,
171
+ "step": 2400
172
+ },
173
+ {
174
+ "epoch": 8.29,
175
+ "learning_rate": 0.00024816291161178504,
176
+ "loss": 0.9997,
177
+ "step": 2500
178
+ },
179
+ {
180
+ "epoch": 8.62,
181
+ "learning_rate": 0.0002455632582322357,
182
+ "loss": 0.995,
183
+ "step": 2600
184
+ },
185
+ {
186
+ "epoch": 8.95,
187
+ "learning_rate": 0.00024296360485268629,
188
+ "loss": 0.9721,
189
+ "step": 2700
190
+ },
191
+ {
192
+ "epoch": 9.28,
193
+ "learning_rate": 0.00024036395147313688,
194
+ "loss": 0.9639,
195
+ "step": 2800
196
+ },
197
+ {
198
+ "epoch": 9.61,
199
+ "learning_rate": 0.0002377642980935875,
200
+ "loss": 0.9543,
201
+ "step": 2900
202
+ },
203
+ {
204
+ "epoch": 9.94,
205
+ "learning_rate": 0.0002351646447140381,
206
+ "loss": 0.9529,
207
+ "step": 3000
208
+ },
209
+ {
210
+ "epoch": 9.94,
211
+ "eval_cer": 0.16591232380706064,
212
+ "eval_loss": 0.6726131439208984,
213
+ "eval_runtime": 144.4701,
214
+ "eval_samples_per_second": 33.495,
215
+ "eval_steps_per_second": 2.097,
216
+ "step": 3000
217
+ },
218
+ {
219
+ "epoch": 10.27,
220
+ "learning_rate": 0.0002325649913344887,
221
+ "loss": 0.9281,
222
+ "step": 3100
223
+ },
224
+ {
225
+ "epoch": 10.6,
226
+ "learning_rate": 0.00022996533795493934,
227
+ "loss": 0.9274,
228
+ "step": 3200
229
+ },
230
+ {
231
+ "epoch": 10.94,
232
+ "learning_rate": 0.00022736568457538994,
233
+ "loss": 0.9149,
234
+ "step": 3300
235
+ },
236
+ {
237
+ "epoch": 11.27,
238
+ "learning_rate": 0.00022476603119584053,
239
+ "loss": 0.9036,
240
+ "step": 3400
241
+ },
242
+ {
243
+ "epoch": 11.6,
244
+ "learning_rate": 0.00022216637781629115,
245
+ "loss": 0.8838,
246
+ "step": 3500
247
+ },
248
+ {
249
+ "epoch": 11.93,
250
+ "learning_rate": 0.00021956672443674175,
251
+ "loss": 0.8872,
252
+ "step": 3600
253
+ },
254
+ {
255
+ "epoch": 12.26,
256
+ "learning_rate": 0.00021696707105719234,
257
+ "loss": 0.8661,
258
+ "step": 3700
259
+ },
260
+ {
261
+ "epoch": 12.59,
262
+ "learning_rate": 0.00021436741767764296,
263
+ "loss": 0.8529,
264
+ "step": 3800
265
+ },
266
+ {
267
+ "epoch": 12.92,
268
+ "learning_rate": 0.00021176776429809356,
269
+ "loss": 0.8566,
270
+ "step": 3900
271
+ },
272
+ {
273
+ "epoch": 13.26,
274
+ "learning_rate": 0.00020916811091854415,
275
+ "loss": 0.8424,
276
+ "step": 4000
277
+ },
278
+ {
279
+ "epoch": 13.26,
280
+ "eval_cer": 0.1511541445751972,
281
+ "eval_loss": 0.6138216257095337,
282
+ "eval_runtime": 143.2958,
283
+ "eval_samples_per_second": 33.769,
284
+ "eval_steps_per_second": 2.115,
285
+ "step": 4000
286
+ },
287
+ {
288
+ "epoch": 13.59,
289
+ "learning_rate": 0.0002065684575389948,
290
+ "loss": 0.8403,
291
+ "step": 4100
292
+ },
293
+ {
294
+ "epoch": 13.92,
295
+ "learning_rate": 0.0002039688041594454,
296
+ "loss": 0.836,
297
+ "step": 4200
298
+ },
299
+ {
300
+ "epoch": 14.25,
301
+ "learning_rate": 0.000201369150779896,
302
+ "loss": 0.8119,
303
+ "step": 4300
304
+ },
305
+ {
306
+ "epoch": 14.58,
307
+ "learning_rate": 0.0001987694974003466,
308
+ "loss": 0.8298,
309
+ "step": 4400
310
+ },
311
+ {
312
+ "epoch": 14.91,
313
+ "learning_rate": 0.0001961698440207972,
314
+ "loss": 0.8019,
315
+ "step": 4500
316
+ },
317
+ {
318
+ "epoch": 15.24,
319
+ "learning_rate": 0.0001935701906412478,
320
+ "loss": 0.8,
321
+ "step": 4600
322
+ },
323
+ {
324
+ "epoch": 15.58,
325
+ "learning_rate": 0.00019097053726169842,
326
+ "loss": 0.7909,
327
+ "step": 4700
328
+ },
329
+ {
330
+ "epoch": 15.91,
331
+ "learning_rate": 0.00018837088388214902,
332
+ "loss": 0.7941,
333
+ "step": 4800
334
+ },
335
+ {
336
+ "epoch": 16.24,
337
+ "learning_rate": 0.0001857712305025996,
338
+ "loss": 0.7888,
339
+ "step": 4900
340
+ },
341
+ {
342
+ "epoch": 16.57,
343
+ "learning_rate": 0.00018317157712305026,
344
+ "loss": 0.767,
345
+ "step": 5000
346
+ },
347
+ {
348
+ "epoch": 16.57,
349
+ "eval_cer": 0.14326587352903142,
350
+ "eval_loss": 0.5884789228439331,
351
+ "eval_runtime": 143.5148,
352
+ "eval_samples_per_second": 33.718,
353
+ "eval_steps_per_second": 2.111,
354
+ "step": 5000
355
+ },
356
+ {
357
+ "epoch": 16.9,
358
+ "learning_rate": 0.00018057192374350086,
359
+ "loss": 0.7655,
360
+ "step": 5100
361
+ },
362
+ {
363
+ "epoch": 17.23,
364
+ "learning_rate": 0.00017797227036395145,
365
+ "loss": 0.7571,
366
+ "step": 5200
367
+ },
368
+ {
369
+ "epoch": 17.56,
370
+ "learning_rate": 0.00017537261698440207,
371
+ "loss": 0.754,
372
+ "step": 5300
373
+ },
374
+ {
375
+ "epoch": 17.9,
376
+ "learning_rate": 0.00017277296360485267,
377
+ "loss": 0.764,
378
+ "step": 5400
379
+ },
380
+ {
381
+ "epoch": 18.23,
382
+ "learning_rate": 0.0001701733102253033,
383
+ "loss": 0.7461,
384
+ "step": 5500
385
+ },
386
+ {
387
+ "epoch": 18.56,
388
+ "learning_rate": 0.00016757365684575388,
389
+ "loss": 0.7363,
390
+ "step": 5600
391
+ },
392
+ {
393
+ "epoch": 18.89,
394
+ "learning_rate": 0.00016497400346620448,
395
+ "loss": 0.7315,
396
+ "step": 5700
397
+ },
398
+ {
399
+ "epoch": 19.22,
400
+ "learning_rate": 0.0001623743500866551,
401
+ "loss": 0.7193,
402
+ "step": 5800
403
+ },
404
+ {
405
+ "epoch": 19.55,
406
+ "learning_rate": 0.00015977469670710572,
407
+ "loss": 0.7253,
408
+ "step": 5900
409
+ },
410
+ {
411
+ "epoch": 19.88,
412
+ "learning_rate": 0.00015717504332755632,
413
+ "loss": 0.7201,
414
+ "step": 6000
415
+ },
416
+ {
417
+ "epoch": 19.88,
418
+ "eval_cer": 0.1377699469804733,
419
+ "eval_loss": 0.5682498812675476,
420
+ "eval_runtime": 145.5022,
421
+ "eval_samples_per_second": 33.257,
422
+ "eval_steps_per_second": 2.082,
423
+ "step": 6000
424
+ },
425
+ {
426
+ "epoch": 20.22,
427
+ "learning_rate": 0.00015457538994800694,
428
+ "loss": 0.7203,
429
+ "step": 6100
430
+ },
431
+ {
432
+ "epoch": 20.55,
433
+ "learning_rate": 0.00015197573656845753,
434
+ "loss": 0.7045,
435
+ "step": 6200
436
+ },
437
+ {
438
+ "epoch": 20.88,
439
+ "learning_rate": 0.00014937608318890813,
440
+ "loss": 0.7072,
441
+ "step": 6300
442
+ },
443
+ {
444
+ "epoch": 21.21,
445
+ "learning_rate": 0.00014677642980935872,
446
+ "loss": 0.6979,
447
+ "step": 6400
448
+ },
449
+ {
450
+ "epoch": 21.54,
451
+ "learning_rate": 0.00014417677642980934,
452
+ "loss": 0.6971,
453
+ "step": 6500
454
+ },
455
+ {
456
+ "epoch": 21.87,
457
+ "learning_rate": 0.00014157712305025997,
458
+ "loss": 0.702,
459
+ "step": 6600
460
+ },
461
+ {
462
+ "epoch": 22.2,
463
+ "learning_rate": 0.00013897746967071056,
464
+ "loss": 0.6767,
465
+ "step": 6700
466
+ },
467
+ {
468
+ "epoch": 22.54,
469
+ "learning_rate": 0.00013637781629116116,
470
+ "loss": 0.682,
471
+ "step": 6800
472
+ },
473
+ {
474
+ "epoch": 22.87,
475
+ "learning_rate": 0.00013377816291161178,
476
+ "loss": 0.6729,
477
+ "step": 6900
478
+ },
479
+ {
480
+ "epoch": 23.2,
481
+ "learning_rate": 0.00013120450606585787,
482
+ "loss": 0.664,
483
+ "step": 7000
484
+ },
485
+ {
486
+ "epoch": 23.2,
487
+ "eval_cer": 0.13325197206776154,
488
+ "eval_loss": 0.5583386421203613,
489
+ "eval_runtime": 143.6376,
490
+ "eval_samples_per_second": 33.689,
491
+ "eval_steps_per_second": 2.109,
492
+ "step": 7000
493
+ },
494
+ {
495
+ "epoch": 23.53,
496
+ "learning_rate": 0.0001286048526863085,
497
+ "loss": 0.668,
498
+ "step": 7100
499
+ },
500
+ {
501
+ "epoch": 23.86,
502
+ "learning_rate": 0.00012600519930675908,
503
+ "loss": 0.67,
504
+ "step": 7200
505
+ },
506
+ {
507
+ "epoch": 24.19,
508
+ "learning_rate": 0.0001234055459272097,
509
+ "loss": 0.6608,
510
+ "step": 7300
511
+ },
512
+ {
513
+ "epoch": 24.52,
514
+ "learning_rate": 0.0001208058925476603,
515
+ "loss": 0.6543,
516
+ "step": 7400
517
+ },
518
+ {
519
+ "epoch": 24.86,
520
+ "learning_rate": 0.0001182062391681109,
521
+ "loss": 0.6648,
522
+ "step": 7500
523
+ },
524
+ {
525
+ "epoch": 25.19,
526
+ "learning_rate": 0.00011560658578856152,
527
+ "loss": 0.6473,
528
+ "step": 7600
529
+ },
530
+ {
531
+ "epoch": 25.52,
532
+ "learning_rate": 0.00011300693240901212,
533
+ "loss": 0.6546,
534
+ "step": 7700
535
+ },
536
+ {
537
+ "epoch": 25.85,
538
+ "learning_rate": 0.00011040727902946273,
539
+ "loss": 0.6397,
540
+ "step": 7800
541
+ },
542
+ {
543
+ "epoch": 26.18,
544
+ "learning_rate": 0.00010780762564991333,
545
+ "loss": 0.6395,
546
+ "step": 7900
547
+ },
548
+ {
549
+ "epoch": 26.51,
550
+ "learning_rate": 0.00010520797227036395,
551
+ "loss": 0.6296,
552
+ "step": 8000
553
+ },
554
+ {
555
+ "epoch": 26.51,
556
+ "eval_cer": 0.1298170179749127,
557
+ "eval_loss": 0.5415648818016052,
558
+ "eval_runtime": 142.9992,
559
+ "eval_samples_per_second": 33.839,
560
+ "eval_steps_per_second": 2.119,
561
+ "step": 8000
562
+ },
563
+ {
564
+ "epoch": 26.84,
565
+ "learning_rate": 0.00010260831889081456,
566
+ "loss": 0.6284,
567
+ "step": 8100
568
+ },
569
+ {
570
+ "epoch": 27.17,
571
+ "learning_rate": 0.00010000866551126515,
572
+ "loss": 0.6328,
573
+ "step": 8200
574
+ },
575
+ {
576
+ "epoch": 27.51,
577
+ "learning_rate": 9.740901213171576e-05,
578
+ "loss": 0.6233,
579
+ "step": 8300
580
+ },
581
+ {
582
+ "epoch": 27.84,
583
+ "learning_rate": 9.480935875216638e-05,
584
+ "loss": 0.6226,
585
+ "step": 8400
586
+ },
587
+ {
588
+ "epoch": 28.17,
589
+ "learning_rate": 9.220970537261698e-05,
590
+ "loss": 0.6214,
591
+ "step": 8500
592
+ },
593
+ {
594
+ "epoch": 28.5,
595
+ "learning_rate": 8.961005199306758e-05,
596
+ "loss": 0.6195,
597
+ "step": 8600
598
+ },
599
+ {
600
+ "epoch": 28.83,
601
+ "learning_rate": 8.701039861351819e-05,
602
+ "loss": 0.6138,
603
+ "step": 8700
604
+ },
605
+ {
606
+ "epoch": 29.16,
607
+ "learning_rate": 8.441074523396879e-05,
608
+ "loss": 0.6067,
609
+ "step": 8800
610
+ },
611
+ {
612
+ "epoch": 29.49,
613
+ "learning_rate": 8.181109185441941e-05,
614
+ "loss": 0.6076,
615
+ "step": 8900
616
+ },
617
+ {
618
+ "epoch": 29.83,
619
+ "learning_rate": 7.921143847487002e-05,
620
+ "loss": 0.6021,
621
+ "step": 9000
622
+ },
623
+ {
624
+ "epoch": 29.83,
625
+ "eval_cer": 0.12716604163972586,
626
+ "eval_loss": 0.5376999974250793,
627
+ "eval_runtime": 142.2332,
628
+ "eval_samples_per_second": 34.022,
629
+ "eval_steps_per_second": 2.13,
630
+ "step": 9000
631
+ },
632
+ {
633
+ "epoch": 30.16,
634
+ "learning_rate": 7.661178509532061e-05,
635
+ "loss": 0.5996,
636
+ "step": 9100
637
+ },
638
+ {
639
+ "epoch": 30.49,
640
+ "learning_rate": 7.401213171577122e-05,
641
+ "loss": 0.5989,
642
+ "step": 9200
643
+ },
644
+ {
645
+ "epoch": 30.82,
646
+ "learning_rate": 7.141247833622183e-05,
647
+ "loss": 0.5924,
648
+ "step": 9300
649
+ },
650
+ {
651
+ "epoch": 31.15,
652
+ "learning_rate": 6.881282495667244e-05,
653
+ "loss": 0.5858,
654
+ "step": 9400
655
+ },
656
+ {
657
+ "epoch": 31.48,
658
+ "learning_rate": 6.621317157712304e-05,
659
+ "loss": 0.5846,
660
+ "step": 9500
661
+ },
662
+ {
663
+ "epoch": 31.81,
664
+ "learning_rate": 6.363951473136915e-05,
665
+ "loss": 0.5906,
666
+ "step": 9600
667
+ },
668
+ {
669
+ "epoch": 32.15,
670
+ "learning_rate": 6.103986135181975e-05,
671
+ "loss": 0.5823,
672
+ "step": 9700
673
+ },
674
+ {
675
+ "epoch": 32.48,
676
+ "learning_rate": 5.8466204506065855e-05,
677
+ "loss": 0.5853,
678
+ "step": 9800
679
+ },
680
+ {
681
+ "epoch": 32.81,
682
+ "learning_rate": 5.586655112651646e-05,
683
+ "loss": 0.5759,
684
+ "step": 9900
685
+ },
686
+ {
687
+ "epoch": 33.14,
688
+ "learning_rate": 5.3266897746967065e-05,
689
+ "loss": 0.568,
690
+ "step": 10000
691
+ },
692
+ {
693
+ "epoch": 33.14,
694
+ "eval_cer": 0.1245797232639338,
695
+ "eval_loss": 0.5240569710731506,
696
+ "eval_runtime": 143.5422,
697
+ "eval_samples_per_second": 33.711,
698
+ "eval_steps_per_second": 2.111,
699
+ "step": 10000
700
+ },
701
+ {
702
+ "epoch": 33.47,
703
+ "learning_rate": 5.066724436741767e-05,
704
+ "loss": 0.574,
705
+ "step": 10100
706
+ },
707
+ {
708
+ "epoch": 33.8,
709
+ "learning_rate": 4.806759098786828e-05,
710
+ "loss": 0.5682,
711
+ "step": 10200
712
+ },
713
+ {
714
+ "epoch": 34.13,
715
+ "learning_rate": 4.546793760831888e-05,
716
+ "loss": 0.5636,
717
+ "step": 10300
718
+ },
719
+ {
720
+ "epoch": 34.47,
721
+ "learning_rate": 4.28682842287695e-05,
722
+ "loss": 0.5651,
723
+ "step": 10400
724
+ },
725
+ {
726
+ "epoch": 34.8,
727
+ "learning_rate": 4.02686308492201e-05,
728
+ "loss": 0.5693,
729
+ "step": 10500
730
+ },
731
+ {
732
+ "epoch": 35.13,
733
+ "learning_rate": 3.766897746967071e-05,
734
+ "loss": 0.5562,
735
+ "step": 10600
736
+ },
737
+ {
738
+ "epoch": 35.46,
739
+ "learning_rate": 3.5069324090121316e-05,
740
+ "loss": 0.5564,
741
+ "step": 10700
742
+ },
743
+ {
744
+ "epoch": 35.79,
745
+ "learning_rate": 3.2469670710571924e-05,
746
+ "loss": 0.5613,
747
+ "step": 10800
748
+ },
749
+ {
750
+ "epoch": 36.12,
751
+ "learning_rate": 2.9870017331022526e-05,
752
+ "loss": 0.5519,
753
+ "step": 10900
754
+ },
755
+ {
756
+ "epoch": 36.45,
757
+ "learning_rate": 2.7270363951473134e-05,
758
+ "loss": 0.5519,
759
+ "step": 11000
760
+ },
761
+ {
762
+ "epoch": 36.45,
763
+ "eval_cer": 0.12275313591103065,
764
+ "eval_loss": 0.5184260606765747,
765
+ "eval_runtime": 142.3486,
766
+ "eval_samples_per_second": 33.994,
767
+ "eval_steps_per_second": 2.129,
768
+ "step": 11000
769
+ },
770
+ {
771
+ "epoch": 36.79,
772
+ "learning_rate": 2.4670710571923742e-05,
773
+ "loss": 0.5605,
774
+ "step": 11100
775
+ },
776
+ {
777
+ "epoch": 37.12,
778
+ "learning_rate": 2.207105719237435e-05,
779
+ "loss": 0.5512,
780
+ "step": 11200
781
+ },
782
+ {
783
+ "epoch": 37.45,
784
+ "learning_rate": 1.9471403812824955e-05,
785
+ "loss": 0.5539,
786
+ "step": 11300
787
+ },
788
+ {
789
+ "epoch": 37.78,
790
+ "learning_rate": 1.687175043327556e-05,
791
+ "loss": 0.5467,
792
+ "step": 11400
793
+ },
794
+ {
795
+ "epoch": 38.11,
796
+ "learning_rate": 1.4272097053726168e-05,
797
+ "loss": 0.5464,
798
+ "step": 11500
799
+ },
800
+ {
801
+ "epoch": 38.44,
802
+ "learning_rate": 1.1672443674176775e-05,
803
+ "loss": 0.5457,
804
+ "step": 11600
805
+ },
806
+ {
807
+ "epoch": 38.77,
808
+ "learning_rate": 9.072790294627383e-06,
809
+ "loss": 0.5445,
810
+ "step": 11700
811
+ },
812
+ {
813
+ "epoch": 39.11,
814
+ "learning_rate": 6.473136915077989e-06,
815
+ "loss": 0.5395,
816
+ "step": 11800
817
+ },
818
+ {
819
+ "epoch": 39.44,
820
+ "learning_rate": 3.8734835355285955e-06,
821
+ "loss": 0.542,
822
+ "step": 11900
823
+ },
824
+ {
825
+ "epoch": 39.77,
826
+ "learning_rate": 1.2738301559792027e-06,
827
+ "loss": 0.5395,
828
+ "step": 12000
829
+ },
830
+ {
831
+ "epoch": 39.77,
832
+ "eval_cer": 0.12272080693133325,
833
+ "eval_loss": 0.5156292915344238,
834
+ "eval_runtime": 142.2326,
835
+ "eval_samples_per_second": 34.022,
836
+ "eval_steps_per_second": 2.13,
837
+ "step": 12000
838
+ },
839
+ {
840
+ "epoch": 39.9,
841
+ "step": 12040,
842
+ "total_flos": 1.9557952498596368e+20,
843
+ "train_loss": 1.365207207163307,
844
+ "train_runtime": 27512.8703,
845
+ "train_samples_per_second": 56.137,
846
+ "train_steps_per_second": 0.438
847
+ }
848
+ ],
849
+ "logging_steps": 100,
850
+ "max_steps": 12040,
851
+ "num_train_epochs": 40,
852
+ "save_steps": 1000,
853
+ "total_flos": 1.9557952498596368e+20,
854
+ "trial_name": null,
855
+ "trial_params": null
856
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:651a4f583ecd7f8a3a536ce2eb86832f1e9a2919291d8f9f5361918ab4c980fe
3
+ size 3963
vocab.json ADDED
@@ -0,0 +1,1432 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "0": 1,
3
+ "1": 2,
4
+ "2": 3,
5
+ "3": 4,
6
+ "4": 5,
7
+ "5": 6,
8
+ "6": 7,
9
+ "7": 8,
10
+ "8": 9,
11
+ "9": 10,
12
+ "[PAD]": 1429,
13
+ "[UNK]": 1428,
14
+ "a": 11,
15
+ "b": 12,
16
+ "c": 13,
17
+ "d": 14,
18
+ "e": 15,
19
+ "f": 16,
20
+ "g": 17,
21
+ "h": 18,
22
+ "i": 19,
23
+ "k": 20,
24
+ "l": 21,
25
+ "m": 22,
26
+ "n": 23,
27
+ "o": 24,
28
+ "p": 25,
29
+ "r": 26,
30
+ "s": 27,
31
+ "t": 28,
32
+ "u": 29,
33
+ "v": 30,
34
+ "w": 31,
35
+ "x": 32,
36
+ "y": 33,
37
+ "z": 34,
38
+ "|": 0,
39
+ "ㄱ": 35,
40
+ "ㄴ": 36,
41
+ "ㄹ": 37,
42
+ "ㅁ": 38,
43
+ "ㅅ": 39,
44
+ "ㅇ": 40,
45
+ "ㅈ": 41,
46
+ "ㅡ": 42,
47
+ "가": 43,
48
+ "각": 44,
49
+ "간": 45,
50
+ "갇": 46,
51
+ "갈": 47,
52
+ "감": 48,
53
+ "갑": 49,
54
+ "값": 50,
55
+ "갓": 51,
56
+ "갔": 52,
57
+ "강": 53,
58
+ "갖": 54,
59
+ "같": 55,
60
+ "갚": 56,
61
+ "갛": 57,
62
+ "개": 58,
63
+ "객": 59,
64
+ "갠": 60,
65
+ "갤": 61,
66
+ "갭": 62,
67
+ "갯": 63,
68
+ "갰": 64,
69
+ "갱": 65,
70
+ "갸": 66,
71
+ "걀": 67,
72
+ "걍": 68,
73
+ "걔": 69,
74
+ "거": 70,
75
+ "걱": 71,
76
+ "건": 72,
77
+ "걷": 73,
78
+ "걸": 74,
79
+ "검": 75,
80
+ "겁": 76,
81
+ "것": 77,
82
+ "겉": 78,
83
+ "게": 79,
84
+ "겐": 80,
85
+ "겔": 81,
86
+ "겠": 82,
87
+ "겨": 83,
88
+ "격": 84,
89
+ "겪": 85,
90
+ "견": 86,
91
+ "결": 87,
92
+ "겸": 88,
93
+ "겹": 89,
94
+ "겼": 90,
95
+ "경": 91,
96
+ "곁": 92,
97
+ "계": 93,
98
+ "고": 94,
99
+ "곡": 95,
100
+ "곤": 96,
101
+ "곧": 97,
102
+ "골": 98,
103
+ "곰": 99,
104
+ "곱": 100,
105
+ "곳": 101,
106
+ "공": 102,
107
+ "과": 103,
108
+ "곽": 104,
109
+ "관": 105,
110
+ "괄": 106,
111
+ "괍": 107,
112
+ "광": 108,
113
+ "괜": 109,
114
+ "괴": 110,
115
+ "굉": 111,
116
+ "교": 112,
117
+ "굥": 113,
118
+ "구": 114,
119
+ "국": 115,
120
+ "군": 116,
121
+ "굳": 117,
122
+ "굴": 118,
123
+ "굵": 119,
124
+ "굶": 120,
125
+ "굽": 121,
126
+ "굿": 122,
127
+ "궁": 123,
128
+ "궈": 124,
129
+ "권": 125,
130
+ "궜": 126,
131
+ "궤": 127,
132
+ "귀": 128,
133
+ "규": 129,
134
+ "균": 130,
135
+ "귤": 131,
136
+ "그": 132,
137
+ "극": 133,
138
+ "근": 134,
139
+ "귿": 135,
140
+ "글": 136,
141
+ "긁": 137,
142
+ "금": 138,
143
+ "급": 139,
144
+ "긋": 140,
145
+ "긍": 141,
146
+ "긎": 142,
147
+ "기": 143,
148
+ "긴": 144,
149
+ "길": 145,
150
+ "김": 146,
151
+ "깁": 147,
152
+ "깃": 148,
153
+ "깊": 149,
154
+ "까": 150,
155
+ "깍": 151,
156
+ "깐": 152,
157
+ "깔": 153,
158
+ "깜": 154,
159
+ "깝": 155,
160
+ "깠": 156,
161
+ "깡": 157,
162
+ "깥": 158,
163
+ "깨": 159,
164
+ "깬": 160,
165
+ "깻": 161,
166
+ "깽": 162,
167
+ "꺄": 163,
168
+ "꺼": 164,
169
+ "꺽": 165,
170
+ "껄": 166,
171
+ "껌": 167,
172
+ "껍": 168,
173
+ "껏": 169,
174
+ "껐": 170,
175
+ "껑": 171,
176
+ "께": 172,
177
+ "껴": 173,
178
+ "꼈": 174,
179
+ "꼐": 175,
180
+ "꼬": 176,
181
+ "꼭": 177,
182
+ "꼰": 178,
183
+ "꼴": 179,
184
+ "꼼": 180,
185
+ "꼽": 181,
186
+ "꽂": 182,
187
+ "꽃": 183,
188
+ "꽈": 184,
189
+ "꽉": 185,
190
+ "꽤": 186,
191
+ "꾸": 187,
192
+ "꾹": 188,
193
+ "꾼": 189,
194
+ "꿀": 190,
195
+ "꿈": 191,
196
+ "꿍": 192,
197
+ "꿔": 193,
198
+ "꿨": 194,
199
+ "꿰": 195,
200
+ "뀌": 196,
201
+ "뀐": 197,
202
+ "뀔": 198,
203
+ "뀝": 199,
204
+ "끄": 200,
205
+ "끈": 201,
206
+ "끊": 202,
207
+ "끌": 203,
208
+ "끓": 204,
209
+ "끔": 205,
210
+ "끕": 206,
211
+ "끗": 207,
212
+ "끝": 208,
213
+ "끼": 209,
214
+ "끽": 210,
215
+ "낀": 211,
216
+ "낄": 212,
217
+ "낌": 213,
218
+ "낑": 214,
219
+ "나": 215,
220
+ "낙": 216,
221
+ "낚": 217,
222
+ "난": 218,
223
+ "날": 219,
224
+ "낡": 220,
225
+ "남": 221,
226
+ "납": 222,
227
+ "낫": 223,
228
+ "났": 224,
229
+ "낭": 225,
230
+ "낮": 226,
231
+ "낯": 227,
232
+ "낱": 228,
233
+ "낳": 229,
234
+ "내": 230,
235
+ "낸": 231,
236
+ "낼": 232,
237
+ "냄": 233,
238
+ "냅": 234,
239
+ "냈": 235,
240
+ "냉": 236,
241
+ "냐": 237,
242
+ "냠": 238,
243
+ "냥": 239,
244
+ "너": 240,
245
+ "넉": 241,
246
+ "넌": 242,
247
+ "널": 243,
248
+ "넓": 244,
249
+ "넘": 245,
250
+ "넛": 246,
251
+ "넣": 247,
252
+ "네": 248,
253
+ "넥": 249,
254
+ "넬": 250,
255
+ "넴": 251,
256
+ "넷": 252,
257
+ "녀": 253,
258
+ "녁": 254,
259
+ "년": 255,
260
+ "념": 256,
261
+ "녔": 257,
262
+ "녕": 258,
263
+ "노": 259,
264
+ "녹": 260,
265
+ "논": 261,
266
+ "놀": 262,
267
+ "놈": 263,
268
+ "농": 264,
269
+ "높": 265,
270
+ "놓": 266,
271
+ "놔": 267,
272
+ "놨": 268,
273
+ "뇌": 269,
274
+ "뇨": 270,
275
+ "누": 271,
276
+ "눅": 272,
277
+ "눈": 273,
278
+ "눌": 274,
279
+ "눔": 275,
280
+ "눕": 276,
281
+ "눗": 277,
282
+ "눠": 278,
283
+ "눴": 279,
284
+ "뉘": 280,
285
+ "뉜": 281,
286
+ "뉩": 282,
287
+ "뉴": 283,
288
+ "느": 284,
289
+ "는": 285,
290
+ "늘": 286,
291
+ "늙": 287,
292
+ "늠": 288,
293
+ "늡": 289,
294
+ "능": 290,
295
+ "늦": 291,
296
+ "늬": 292,
297
+ "니": 293,
298
+ "닉": 294,
299
+ "닌": 295,
300
+ "닐": 296,
301
+ "님": 297,
302
+ "닙": 298,
303
+ "닝": 299,
304
+ "다": 300,
305
+ "닥": 301,
306
+ "닦": 302,
307
+ "단": 303,
308
+ "닫": 304,
309
+ "달": 305,
310
+ "닭": 306,
311
+ "닮": 307,
312
+ "담": 308,
313
+ "답": 309,
314
+ "닷": 310,
315
+ "당": 311,
316
+ "닿": 312,
317
+ "대": 313,
318
+ "댁": 314,
319
+ "댄": 315,
320
+ "댈": 316,
321
+ "댐": 317,
322
+ "댑": 318,
323
+ "댓": 319,
324
+ "댔": 320,
325
+ "댕": 321,
326
+ "댜": 322,
327
+ "더": 323,
328
+ "덕": 324,
329
+ "던": 325,
330
+ "덜": 326,
331
+ "덟": 327,
332
+ "덤": 328,
333
+ "덥": 329,
334
+ "덧": 330,
335
+ "덩": 331,
336
+ "덮": 332,
337
+ "데": 333,
338
+ "덱": 334,
339
+ "델": 335,
340
+ "뎀": 336,
341
+ "뎅": 337,
342
+ "뎌": 338,
343
+ "도": 339,
344
+ "독": 340,
345
+ "돈": 341,
346
+ "돋": 342,
347
+ "돌": 343,
348
+ "돕": 344,
349
+ "돗": 345,
350
+ "동": 346,
351
+ "돼": 347,
352
+ "됐": 348,
353
+ "되": 349,
354
+ "된": 350,
355
+ "될": 351,
356
+ "됨": 352,
357
+ "됩": 353,
358
+ "됬": 354,
359
+ "두": 355,
360
+ "둑": 356,
361
+ "둔": 357,
362
+ "둘": 358,
363
+ "둠": 359,
364
+ "둡": 360,
365
+ "둥": 361,
366
+ "둬": 362,
367
+ "뒀": 363,
368
+ "뒤": 364,
369
+ "뒷": 365,
370
+ "듀": 366,
371
+ "드": 367,
372
+ "득": 368,
373
+ "든": 369,
374
+ "듣": 370,
375
+ "들": 371,
376
+ "듦": 372,
377
+ "듬": 373,
378
+ "듭": 374,
379
+ "듯": 375,
380
+ "등": 376,
381
+ "디": 377,
382
+ "딘": 378,
383
+ "딜": 379,
384
+ "딤": 380,
385
+ "딥": 381,
386
+ "딨": 382,
387
+ "딩": 383,
388
+ "딪": 384,
389
+ "따": 385,
390
+ "딱": 386,
391
+ "딴": 387,
392
+ "딸": 388,
393
+ "땀": 389,
394
+ "땅": 390,
395
+ "때": 391,
396
+ "땐": 392,
397
+ "땜": 393,
398
+ "땠": 394,
399
+ "땡": 395,
400
+ "떄": 396,
401
+ "떠": 397,
402
+ "떡": 398,
403
+ "떤": 399,
404
+ "떨": 400,
405
+ "떴": 401,
406
+ "떵": 402,
407
+ "떻": 403,
408
+ "떼": 404,
409
+ "뗀": 405,
410
+ "뗐": 406,
411
+ "또": 407,
412
+ "똑": 408,
413
+ "똠": 409,
414
+ "똥": 410,
415
+ "뚜": 411,
416
+ "뚝": 412,
417
+ "뚤": 413,
418
+ "뚫": 414,
419
+ "뚱": 415,
420
+ "뛰": 416,
421
+ "뛸": 417,
422
+ "뜨": 418,
423
+ "뜩": 419,
424
+ "뜬": 420,
425
+ "뜯": 421,
426
+ "뜰": 422,
427
+ "뜸": 423,
428
+ "뜹": 424,
429
+ "뜻": 425,
430
+ "띄": 426,
431
+ "띈": 427,
432
+ "띌": 428,
433
+ "띕": 429,
434
+ "띠": 430,
435
+ "띡": 431,
436
+ "띵": 432,
437
+ "라": 433,
438
+ "락": 434,
439
+ "란": 435,
440
+ "랄": 436,
441
+ "람": 437,
442
+ "랍": 438,
443
+ "랏": 439,
444
+ "랐": 440,
445
+ "랑": 441,
446
+ "랗": 442,
447
+ "래": 443,
448
+ "랙": 444,
449
+ "랜": 445,
450
+ "램": 446,
451
+ "랩": 447,
452
+ "랫": 448,
453
+ "랬": 449,
454
+ "략": 450,
455
+ "럈": 451,
456
+ "량": 452,
457
+ "러": 453,
458
+ "럭": 454,
459
+ "런": 455,
460
+ "럴": 456,
461
+ "럼": 457,
462
+ "럽": 458,
463
+ "렀": 459,
464
+ "렁": 460,
465
+ "렇": 461,
466
+ "레": 462,
467
+ "렉": 463,
468
+ "렌": 464,
469
+ "렘": 465,
470
+ "렛": 466,
471
+ "려": 467,
472
+ "력": 468,
473
+ "련": 469,
474
+ "렬": 470,
475
+ "렴": 471,
476
+ "렵": 472,
477
+ "렷": 473,
478
+ "렸": 474,
479
+ "령": 475,
480
+ "례": 476,
481
+ "로": 477,
482
+ "록": 478,
483
+ "론": 479,
484
+ "롤": 480,
485
+ "롬": 481,
486
+ "롭": 482,
487
+ "롯": 483,
488
+ "롱": 484,
489
+ "롷": 485,
490
+ "뢰": 486,
491
+ "료": 487,
492
+ "룡": 488,
493
+ "루": 489,
494
+ "룩": 490,
495
+ "룬": 491,
496
+ "룰": 492,
497
+ "룹": 493,
498
+ "뤄": 494,
499
+ "뤘": 495,
500
+ "류": 496,
501
+ "률": 497,
502
+ "륨": 498,
503
+ "륭": 499,
504
+ "르": 500,
505
+ "륵": 501,
506
+ "른": 502,
507
+ "를": 503,
508
+ "름": 504,
509
+ "릅": 505,
510
+ "릇": 506,
511
+ "릉": 507,
512
+ "릎": 508,
513
+ "리": 509,
514
+ "릭": 510,
515
+ "린": 511,
516
+ "릴": 512,
517
+ "림": 513,
518
+ "립": 514,
519
+ "릿": 515,
520
+ "링": 516,
521
+ "마": 517,
522
+ "막": 518,
523
+ "만": 519,
524
+ "많": 520,
525
+ "말": 521,
526
+ "맑": 522,
527
+ "맘": 523,
528
+ "맙": 524,
529
+ "맛": 525,
530
+ "망": 526,
531
+ "맞": 527,
532
+ "맡": 528,
533
+ "맣": 529,
534
+ "매": 530,
535
+ "맥": 531,
536
+ "맨": 532,
537
+ "맴": 533,
538
+ "맵": 534,
539
+ "맷": 535,
540
+ "맸": 536,
541
+ "맹": 537,
542
+ "맺": 538,
543
+ "머": 539,
544
+ "먹": 540,
545
+ "먼": 541,
546
+ "멀": 542,
547
+ "멈": 543,
548
+ "멋": 544,
549
+ "멍": 545,
550
+ "메": 546,
551
+ "멕": 547,
552
+ "멘": 548,
553
+ "멜": 549,
554
+ "멤": 550,
555
+ "멧": 551,
556
+ "며": 552,
557
+ "면": 553,
558
+ "멸": 554,
559
+ "명": 555,
560
+ "몇": 556,
561
+ "모": 557,
562
+ "목": 558,
563
+ "몫": 559,
564
+ "몬": 560,
565
+ "몰": 561,
566
+ "몸": 562,
567
+ "몹": 563,
568
+ "못": 564,
569
+ "몽": 565,
570
+ "묘": 566,
571
+ "무": 567,
572
+ "묵": 568,
573
+ "묶": 569,
574
+ "문": 570,
575
+ "묻": 571,
576
+ "물": 572,
577
+ "묽": 573,
578
+ "뭇": 574,
579
+ "뭉": 575,
580
+ "뭐": 576,
581
+ "뭔": 577,
582
+ "뭘": 578,
583
+ "뭡": 579,
584
+ "뭣": 580,
585
+ "뮤": 581,
586
+ "므": 582,
587
+ "믄": 583,
588
+ "믈": 584,
589
+ "미": 585,
590
+ "믹": 586,
591
+ "민": 587,
592
+ "믿": 588,
593
+ "밀": 589,
594
+ "밈": 590,
595
+ "밌": 591,
596
+ "밍": 592,
597
+ "및": 593,
598
+ "밑": 594,
599
+ "바": 595,
600
+ "박": 596,
601
+ "밖": 597,
602
+ "반": 598,
603
+ "받": 599,
604
+ "발": 600,
605
+ "밝": 601,
606
+ "밟": 602,
607
+ "밤": 603,
608
+ "밥": 604,
609
+ "방": 605,
610
+ "배": 606,
611
+ "백": 607,
612
+ "밴": 608,
613
+ "밸": 609,
614
+ "뱁": 610,
615
+ "뱃": 611,
616
+ "뱅": 612,
617
+ "뱉": 613,
618
+ "버": 614,
619
+ "벅": 615,
620
+ "번": 616,
621
+ "벌": 617,
622
+ "범": 618,
623
+ "법": 619,
624
+ "벗": 620,
625
+ "벙": 621,
626
+ "벚": 622,
627
+ "베": 623,
628
+ "벡": 624,
629
+ "벤": 625,
630
+ "벨": 626,
631
+ "벳": 627,
632
+ "벼": 628,
633
+ "벽": 629,
634
+ "변": 630,
635
+ "별": 631,
636
+ "볍": 632,
637
+ "병": 633,
638
+ "볕": 634,
639
+ "보": 635,
640
+ "복": 636,
641
+ "볶": 637,
642
+ "본": 638,
643
+ "볼": 639,
644
+ "봄": 640,
645
+ "봅": 641,
646
+ "봈": 642,
647
+ "봉": 643,
648
+ "봐": 644,
649
+ "봤": 645,
650
+ "봬": 646,
651
+ "뵀": 647,
652
+ "뵌": 648,
653
+ "뵐": 649,
654
+ "뵙": 650,
655
+ "부": 651,
656
+ "북": 652,
657
+ "분": 653,
658
+ "붇": 654,
659
+ "불": 655,
660
+ "붉": 656,
661
+ "붓": 657,
662
+ "붙": 658,
663
+ "뷔": 659,
664
+ "뷰": 660,
665
+ "브": 661,
666
+ "븐": 662,
667
+ "블": 663,
668
+ "비": 664,
669
+ "빅": 665,
670
+ "빈": 666,
671
+ "빌": 667,
672
+ "빔": 668,
673
+ "빗": 669,
674
+ "빙": 670,
675
+ "빛": 671,
676
+ "빠": 672,
677
+ "빡": 673,
678
+ "빤": 674,
679
+ "빨": 675,
680
+ "빴": 676,
681
+ "빵": 677,
682
+ "빼": 678,
683
+ "빽": 679,
684
+ "뺀": 680,
685
+ "뺄": 681,
686
+ "뺏": 682,
687
+ "뺐": 683,
688
+ "뺑": 684,
689
+ "뺨": 685,
690
+ "뺭": 686,
691
+ "뻐": 687,
692
+ "뻑": 688,
693
+ "뻔": 689,
694
+ "뻗": 690,
695
+ "뻘": 691,
696
+ "뻣": 692,
697
+ "뻤": 693,
698
+ "뻥": 694,
699
+ "뼈": 695,
700
+ "뽀": 696,
701
+ "뽑": 697,
702
+ "뽕": 698,
703
+ "뾰": 699,
704
+ "뿅": 700,
705
+ "뿌": 701,
706
+ "뿍": 702,
707
+ "뿐": 703,
708
+ "뿔": 704,
709
+ "쁘": 705,
710
+ "쁜": 706,
711
+ "쁠": 707,
712
+ "쁨": 708,
713
+ "쁩": 709,
714
+ "삐": 710,
715
+ "삔": 711,
716
+ "삥": 712,
717
+ "사": 713,
718
+ "삭": 714,
719
+ "산": 715,
720
+ "살": 716,
721
+ "삶": 717,
722
+ "삼": 718,
723
+ "삽": 719,
724
+ "삿": 720,
725
+ "샀": 721,
726
+ "상": 722,
727
+ "새": 723,
728
+ "색": 724,
729
+ "샌": 725,
730
+ "샘": 726,
731
+ "생": 727,
732
+ "샤": 728,
733
+ "샥": 729,
734
+ "샴": 730,
735
+ "샵": 731,
736
+ "서": 732,
737
+ "석": 733,
738
+ "섞": 734,
739
+ "선": 735,
740
+ "설": 736,
741
+ "섬": 737,
742
+ "섭": 738,
743
+ "섯": 739,
744
+ "섰": 740,
745
+ "성": 741,
746
+ "세": 742,
747
+ "섹": 743,
748
+ "센": 744,
749
+ "셀": 745,
750
+ "셈": 746,
751
+ "셉": 747,
752
+ "셋": 748,
753
+ "셌": 749,
754
+ "셔": 750,
755
+ "션": 751,
756
+ "셜": 752,
757
+ "셨": 753,
758
+ "셰": 754,
759
+ "소": 755,
760
+ "속": 756,
761
+ "손": 757,
762
+ "솔": 758,
763
+ "솜": 759,
764
+ "솟": 760,
765
+ "송": 761,
766
+ "솥": 762,
767
+ "쇄": 763,
768
+ "쇠": 764,
769
+ "쇼": 765,
770
+ "숀": 766,
771
+ "숍": 767,
772
+ "수": 768,
773
+ "숙": 769,
774
+ "순": 770,
775
+ "숟": 771,
776
+ "술": 772,
777
+ "숨": 773,
778
+ "숫": 774,
779
+ "숭": 775,
780
+ "숯": 776,
781
+ "숱": 777,
782
+ "숲": 778,
783
+ "숴": 779,
784
+ "쉐": 780,
785
+ "쉬": 781,
786
+ "쉴": 782,
787
+ "쉽": 783,
788
+ "슈": 784,
789
+ "슉": 785,
790
+ "슘": 786,
791
+ "슛": 787,
792
+ "슝": 788,
793
+ "스": 789,
794
+ "슥": 790,
795
+ "슨": 791,
796
+ "슬": 792,
797
+ "슴": 793,
798
+ "습": 794,
799
+ "슷": 795,
800
+ "승": 796,
801
+ "시": 797,
802
+ "식": 798,
803
+ "신": 799,
804
+ "싣": 800,
805
+ "실": 801,
806
+ "싫": 802,
807
+ "심": 803,
808
+ "십": 804,
809
+ "싱": 805,
810
+ "싴": 806,
811
+ "싶": 807,
812
+ "싸": 808,
813
+ "싹": 809,
814
+ "싼": 810,
815
+ "쌀": 811,
816
+ "쌈": 812,
817
+ "쌉": 813,
818
+ "쌍": 814,
819
+ "쌓": 815,
820
+ "쌤": 816,
821
+ "쌩": 817,
822
+ "쌰": 818,
823
+ "썜": 819,
824
+ "써": 820,
825
+ "썩": 821,
826
+ "썬": 822,
827
+ "썰": 823,
828
+ "썸": 824,
829
+ "썹": 825,
830
+ "썼": 826,
831
+ "썽": 827,
832
+ "쎄": 828,
833
+ "쎈": 829,
834
+ "쎼": 830,
835
+ "쏘": 831,
836
+ "쏙": 832,
837
+ "쏜": 833,
838
+ "쏟": 834,
839
+ "쏠": 835,
840
+ "쏴": 836,
841
+ "쑉": 837,
842
+ "쑤": 838,
843
+ "쑥": 839,
844
+ "쓰": 840,
845
+ "쓱": 841,
846
+ "쓴": 842,
847
+ "쓸": 843,
848
+ "씀": 844,
849
+ "씁": 845,
850
+ "씌": 846,
851
+ "씨": 847,
852
+ "씩": 848,
853
+ "씬": 849,
854
+ "씹": 850,
855
+ "씻": 851,
856
+ "아": 852,
857
+ "악": 853,
858
+ "안": 854,
859
+ "앉": 855,
860
+ "않": 856,
861
+ "알": 857,
862
+ "앓": 858,
863
+ "암": 859,
864
+ "압": 860,
865
+ "앗": 861,
866
+ "았": 862,
867
+ "앙": 863,
868
+ "앞": 864,
869
+ "애": 865,
870
+ "액": 866,
871
+ "앤": 867,
872
+ "앨": 868,
873
+ "앰": 869,
874
+ "앱": 870,
875
+ "앴": 871,
876
+ "앵": 872,
877
+ "야": 873,
878
+ "약": 874,
879
+ "얀": 875,
880
+ "얄": 876,
881
+ "얇": 877,
882
+ "얌": 878,
883
+ "양": 879,
884
+ "얕": 880,
885
+ "얗": 881,
886
+ "얘": 882,
887
+ "얜": 883,
888
+ "어": 884,
889
+ "억": 885,
890
+ "언": 886,
891
+ "얹": 887,
892
+ "얻": 888,
893
+ "얼": 889,
894
+ "엄": 890,
895
+ "업": 891,
896
+ "없": 892,
897
+ "엇": 893,
898
+ "었": 894,
899
+ "엉": 895,
900
+ "엎": 896,
901
+ "에": 897,
902
+ "엑": 898,
903
+ "엔": 899,
904
+ "엘": 900,
905
+ "엠": 901,
906
+ "엥": 902,
907
+ "여": 903,
908
+ "역": 904,
909
+ "연": 905,
910
+ "열": 906,
911
+ "염": 907,
912
+ "엽": 908,
913
+ "엿": 909,
914
+ "였": 910,
915
+ "영": 911,
916
+ "옆": 912,
917
+ "옇": 913,
918
+ "예": 914,
919
+ "옛": 915,
920
+ "오": 916,
921
+ "옥": 917,
922
+ "온": 918,
923
+ "올": 919,
924
+ "옮": 920,
925
+ "옳": 921,
926
+ "옴": 922,
927
+ "옵": 923,
928
+ "옷": 924,
929
+ "옹": 925,
930
+ "와": 926,
931
+ "완": 927,
932
+ "왈": 928,
933
+ "왓": 929,
934
+ "왔": 930,
935
+ "왕": 931,
936
+ "왜": 932,
937
+ "왠": 933,
938
+ "외": 934,
939
+ "왼": 935,
940
+ "요": 936,
941
+ "욕": 937,
942
+ "용": 938,
943
+ "우": 939,
944
+ "욱": 940,
945
+ "운": 941,
946
+ "울": 942,
947
+ "움": 943,
948
+ "웁": 944,
949
+ "웃": 945,
950
+ "웅": 946,
951
+ "워": 947,
952
+ "원": 948,
953
+ "월": 949,
954
+ "웠": 950,
955
+ "웨": 951,
956
+ "웩": 952,
957
+ "웬": 953,
958
+ "웰": 954,
959
+ "웹": 955,
960
+ "위": 956,
961
+ "윈": 957,
962
+ "윗": 958,
963
+ "유": 959,
964
+ "육": 960,
965
+ "윤": 961,
966
+ "율": 962,
967
+ "융": 963,
968
+ "으": 964,
969
+ "윽": 965,
970
+ "은": 966,
971
+ "읃": 967,
972
+ "을": 968,
973
+ "음": 969,
974
+ "읍": 970,
975
+ "읏": 971,
976
+ "응": 972,
977
+ "의": 973,
978
+ "이": 974,
979
+ "익": 975,
980
+ "인": 976,
981
+ "일": 977,
982
+ "읽": 978,
983
+ "잃": 979,
984
+ "임": 980,
985
+ "입": 981,
986
+ "잇": 982,
987
+ "있": 983,
988
+ "잉": 984,
989
+ "잊": 985,
990
+ "잎": 986,
991
+ "자": 987,
992
+ "작": 988,
993
+ "잔": 989,
994
+ "잖": 990,
995
+ "잘": 991,
996
+ "잠": 992,
997
+ "잡": 993,
998
+ "잣": 994,
999
+ "잤": 995,
1000
+ "장": 996,
1001
+ "잦": 997,
1002
+ "재": 998,
1003
+ "잰": 999,
1004
+ "잼": 1000,
1005
+ "잽": 1001,
1006
+ "쟁": 1002,
1007
+ "쟈": 1003,
1008
+ "쟤": 1004,
1009
+ "저": 1005,
1010
+ "적": 1006,
1011
+ "전": 1007,
1012
+ "절": 1008,
1013
+ "젊": 1009,
1014
+ "점": 1010,
1015
+ "접": 1011,
1016
+ "젓": 1012,
1017
+ "정": 1013,
1018
+ "젖": 1014,
1019
+ "제": 1015,
1020
+ "젝": 1016,
1021
+ "젠": 1017,
1022
+ "젤": 1018,
1023
+ "젯": 1019,
1024
+ "져": 1020,
1025
+ "졌": 1021,
1026
+ "조": 1022,
1027
+ "족": 1023,
1028
+ "존": 1024,
1029
+ "졸": 1025,
1030
+ "좀": 1026,
1031
+ "좁": 1027,
1032
+ "종": 1028,
1033
+ "좋": 1029,
1034
+ "좌": 1030,
1035
+ "좟": 1031,
1036
+ "죄": 1032,
1037
+ "죠": 1033,
1038
+ "죵": 1034,
1039
+ "주": 1035,
1040
+ "죽": 1036,
1041
+ "준": 1037,
1042
+ "줄": 1038,
1043
+ "줌": 1039,
1044
+ "줍": 1040,
1045
+ "중": 1041,
1046
+ "줘": 1042,
1047
+ "줬": 1043,
1048
+ "쥐": 1044,
1049
+ "쥬": 1045,
1050
+ "즈": 1046,
1051
+ "즉": 1047,
1052
+ "즌": 1048,
1053
+ "즐": 1049,
1054
+ "즘": 1050,
1055
+ "증": 1051,
1056
+ "지": 1052,
1057
+ "직": 1053,
1058
+ "진": 1054,
1059
+ "질": 1055,
1060
+ "짐": 1056,
1061
+ "집": 1057,
1062
+ "짓": 1058,
1063
+ "징": 1059,
1064
+ "짖": 1060,
1065
+ "짙": 1061,
1066
+ "짚": 1062,
1067
+ "짜": 1063,
1068
+ "짝": 1064,
1069
+ "짠": 1065,
1070
+ "짤": 1066,
1071
+ "짧": 1067,
1072
+ "짬": 1068,
1073
+ "짭": 1069,
1074
+ "짰": 1070,
1075
+ "짱": 1071,
1076
+ "째": 1072,
1077
+ "쨋": 1073,
1078
+ "쨌": 1074,
1079
+ "쨍": 1075,
1080
+ "쨰": 1076,
1081
+ "쩌": 1077,
1082
+ "쩍": 1078,
1083
+ "쩐": 1079,
1084
+ "쩔": 1080,
1085
+ "쩜": 1081,
1086
+ "쩝": 1082,
1087
+ "쩠": 1083,
1088
+ "쩡": 1084,
1089
+ "쪼": 1085,
1090
+ "쪽": 1086,
1091
+ "쫀": 1087,
1092
+ "쫄": 1088,
1093
+ "쫌": 1089,
1094
+ "쫓": 1090,
1095
+ "쫙": 1091,
1096
+ "쫜": 1092,
1097
+ "쬐": 1093,
1098
+ "쭈": 1094,
1099
+ "쭉": 1095,
1100
+ "쭤": 1096,
1101
+ "쮸": 1097,
1102
+ "쯔": 1098,
1103
+ "쯤": 1099,
1104
+ "찌": 1100,
1105
+ "찍": 1101,
1106
+ "찐": 1102,
1107
+ "찔": 1103,
1108
+ "찜": 1104,
1109
+ "찝": 1105,
1110
+ "찡": 1106,
1111
+ "찢": 1107,
1112
+ "차": 1108,
1113
+ "착": 1109,
1114
+ "찬": 1110,
1115
+ "찮": 1111,
1116
+ "찰": 1112,
1117
+ "참": 1113,
1118
+ "찻": 1114,
1119
+ "찼": 1115,
1120
+ "창": 1116,
1121
+ "찾": 1117,
1122
+ "채": 1118,
1123
+ "책": 1119,
1124
+ "챌": 1120,
1125
+ "챕": 1121,
1126
+ "챙": 1122,
1127
+ "처": 1123,
1128
+ "척": 1124,
1129
+ "천": 1125,
1130
+ "철": 1126,
1131
+ "첨": 1127,
1132
+ "첩": 1128,
1133
+ "첫": 1129,
1134
+ "청": 1130,
1135
+ "체": 1131,
1136
+ "첼": 1132,
1137
+ "쳅": 1133,
1138
+ "쳐": 1134,
1139
+ "쳤": 1135,
1140
+ "쳬": 1136,
1141
+ "초": 1137,
1142
+ "촉": 1138,
1143
+ "촌": 1139,
1144
+ "촘": 1140,
1145
+ "촛": 1141,
1146
+ "총": 1142,
1147
+ "촤": 1143,
1148
+ "촥": 1144,
1149
+ "촬": 1145,
1150
+ "최": 1146,
1151
+ "쵸": 1147,
1152
+ "추": 1148,
1153
+ "축": 1149,
1154
+ "춘": 1150,
1155
+ "출": 1151,
1156
+ "춤": 1152,
1157
+ "춥": 1153,
1158
+ "춧": 1154,
1159
+ "충": 1155,
1160
+ "춰": 1156,
1161
+ "췄": 1157,
1162
+ "췌": 1158,
1163
+ "취": 1159,
1164
+ "츄": 1160,
1165
+ "츠": 1161,
1166
+ "측": 1162,
1167
+ "츨": 1163,
1168
+ "층": 1164,
1169
+ "치": 1165,
1170
+ "칙": 1166,
1171
+ "친": 1167,
1172
+ "칠": 1168,
1173
+ "침": 1169,
1174
+ "칩": 1170,
1175
+ "칫": 1171,
1176
+ "칭": 1172,
1177
+ "카": 1173,
1178
+ "칵": 1174,
1179
+ "칸": 1175,
1180
+ "칼": 1176,
1181
+ "캐": 1177,
1182
+ "캔": 1178,
1183
+ "캘": 1179,
1184
+ "캠": 1180,
1185
+ "캡": 1181,
1186
+ "캣": 1182,
1187
+ "캬": 1183,
1188
+ "커": 1184,
1189
+ "컥": 1185,
1190
+ "컨": 1186,
1191
+ "컬": 1187,
1192
+ "컴": 1188,
1193
+ "컵": 1189,
1194
+ "컷": 1190,
1195
+ "컸": 1191,
1196
+ "케": 1192,
1197
+ "켄": 1193,
1198
+ "켓": 1194,
1199
+ "켜": 1195,
1200
+ "켠": 1196,
1201
+ "켰": 1197,
1202
+ "코": 1198,
1203
+ "콕": 1199,
1204
+ "콘": 1200,
1205
+ "콜": 1201,
1206
+ "콤": 1202,
1207
+ "콥": 1203,
1208
+ "콧": 1204,
1209
+ "콩": 1205,
1210
+ "콸": 1206,
1211
+ "쾌": 1207,
1212
+ "쿠": 1208,
1213
+ "쿡": 1209,
1214
+ "쿤": 1210,
1215
+ "쿨": 1211,
1216
+ "쿵": 1212,
1217
+ "퀄": 1213,
1218
+ "퀴": 1214,
1219
+ "큐": 1215,
1220
+ "큘": 1216,
1221
+ "크": 1217,
1222
+ "큰": 1218,
1223
+ "클": 1219,
1224
+ "큼": 1220,
1225
+ "큽": 1221,
1226
+ "키": 1222,
1227
+ "킨": 1223,
1228
+ "킬": 1224,
1229
+ "킴": 1225,
1230
+ "킵": 1226,
1231
+ "킹": 1227,
1232
+ "타": 1228,
1233
+ "탁": 1229,
1234
+ "탄": 1230,
1235
+ "탈": 1231,
1236
+ "탐": 1232,
1237
+ "탑": 1233,
1238
+ "탓": 1234,
1239
+ "탔": 1235,
1240
+ "탕": 1236,
1241
+ "태": 1237,
1242
+ "택": 1238,
1243
+ "탠": 1239,
1244
+ "탱": 1240,
1245
+ "터": 1241,
1246
+ "턱": 1242,
1247
+ "턴": 1243,
1248
+ "털": 1244,
1249
+ "텁": 1245,
1250
+ "텅": 1246,
1251
+ "테": 1247,
1252
+ "텍": 1248,
1253
+ "텐": 1249,
1254
+ "텔": 1250,
1255
+ "템": 1251,
1256
+ "텝": 1252,
1257
+ "텼": 1253,
1258
+ "토": 1254,
1259
+ "톡": 1255,
1260
+ "톤": 1256,
1261
+ "톨": 1257,
1262
+ "톱": 1258,
1263
+ "통": 1259,
1264
+ "퇴": 1260,
1265
+ "투": 1261,
1266
+ "툭": 1262,
1267
+ "툴": 1263,
1268
+ "툼": 1264,
1269
+ "퉁": 1265,
1270
+ "튀": 1266,
1271
+ "튜": 1267,
1272
+ "트": 1268,
1273
+ "특": 1269,
1274
+ "튼": 1270,
1275
+ "틀": 1271,
1276
+ "틈": 1272,
1277
+ "틍": 1273,
1278
+ "티": 1274,
1279
+ "틱": 1275,
1280
+ "틴": 1276,
1281
+ "틸": 1277,
1282
+ "팀": 1278,
1283
+ "팁": 1279,
1284
+ "팅": 1280,
1285
+ "파": 1281,
1286
+ "팍": 1282,
1287
+ "팎": 1283,
1288
+ "판": 1284,
1289
+ "팔": 1285,
1290
+ "팠": 1286,
1291
+ "팡": 1287,
1292
+ "팥": 1288,
1293
+ "패": 1289,
1294
+ "팩": 1290,
1295
+ "팬": 1291,
1296
+ "팽": 1292,
1297
+ "퍼": 1293,
1298
+ "퍽": 1294,
1299
+ "펀": 1295,
1300
+ "펄": 1296,
1301
+ "펌": 1297,
1302
+ "페": 1298,
1303
+ "펙": 1299,
1304
+ "펜": 1300,
1305
+ "펠": 1301,
1306
+ "펩": 1302,
1307
+ "펴": 1303,
1308
+ "편": 1304,
1309
+ "펼": 1305,
1310
+ "폈": 1306,
1311
+ "평": 1307,
1312
+ "폐": 1308,
1313
+ "포": 1309,
1314
+ "폭": 1310,
1315
+ "폰": 1311,
1316
+ "폴": 1312,
1317
+ "폼": 1313,
1318
+ "퐁": 1314,
1319
+ "표": 1315,
1320
+ "푸": 1316,
1321
+ "푹": 1317,
1322
+ "푼": 1318,
1323
+ "풀": 1319,
1324
+ "품": 1320,
1325
+ "풋": 1321,
1326
+ "풍": 1322,
1327
+ "퓨": 1323,
1328
+ "프": 1324,
1329
+ "픈": 1325,
1330
+ "플": 1326,
1331
+ "픔": 1327,
1332
+ "피": 1328,
1333
+ "픽": 1329,
1334
+ "핀": 1330,
1335
+ "필": 1331,
1336
+ "핍": 1332,
1337
+ "핏": 1333,
1338
+ "핑": 1334,
1339
+ "하": 1335,
1340
+ "학": 1336,
1341
+ "한": 1337,
1342
+ "할": 1338,
1343
+ "함": 1339,
1344
+ "합": 1340,
1345
+ "핫": 1341,
1346
+ "항": 1342,
1347
+ "해": 1343,
1348
+ "핵": 1344,
1349
+ "핸": 1345,
1350
+ "핼": 1346,
1351
+ "햄": 1347,
1352
+ "햅": 1348,
1353
+ "햇": 1349,
1354
+ "했": 1350,
1355
+ "행": 1351,
1356
+ "햍": 1352,
1357
+ "햐": 1353,
1358
+ "향": 1354,
1359
+ "헀": 1355,
1360
+ "허": 1356,
1361
+ "헉": 1357,
1362
+ "헌": 1358,
1363
+ "헐": 1359,
1364
+ "험": 1360,
1365
+ "헛": 1361,
1366
+ "헝": 1362,
1367
+ "헤": 1363,
1368
+ "헨": 1364,
1369
+ "헬": 1365,
1370
+ "헷": 1366,
1371
+ "헹": 1367,
1372
+ "혀": 1368,
1373
+ "혁": 1369,
1374
+ "현": 1370,
1375
+ "혈": 1371,
1376
+ "혐": 1372,
1377
+ "협": 1373,
1378
+ "혓": 1374,
1379
+ "혔": 1375,
1380
+ "형": 1376,
1381
+ "혜": 1377,
1382
+ "호": 1378,
1383
+ "혹": 1379,
1384
+ "혼": 1380,
1385
+ "홀": 1381,
1386
+ "홈": 1382,
1387
+ "홉": 1383,
1388
+ "홍": 1384,
1389
+ "화": 1385,
1390
+ "확": 1386,
1391
+ "환": 1387,
1392
+ "활": 1388,
1393
+ "홥": 1389,
1394
+ "황": 1390,
1395
+ "회": 1391,
1396
+ "획": 1392,
1397
+ "횟": 1393,
1398
+ "횡": 1394,
1399
+ "효": 1395,
1400
+ "횩": 1396,
1401
+ "후": 1397,
1402
+ "훅": 1398,
1403
+ "훈": 1399,
1404
+ "훌": 1400,
1405
+ "훑": 1401,
1406
+ "훔": 1402,
1407
+ "훨": 1403,
1408
+ "훼": 1404,
1409
+ "휑": 1405,
1410
+ "휘": 1406,
1411
+ "휙": 1407,
1412
+ "휩": 1408,
1413
+ "휴": 1409,
1414
+ "흉": 1410,
1415
+ "흐": 1411,
1416
+ "흑": 1412,
1417
+ "흔": 1413,
1418
+ "흘": 1414,
1419
+ "흙": 1415,
1420
+ "흠": 1416,
1421
+ "흡": 1417,
1422
+ "흣": 1418,
1423
+ "흥": 1419,
1424
+ "흩": 1420,
1425
+ "희": 1421,
1426
+ "흰": 1422,
1427
+ "히": 1423,
1428
+ "힌": 1424,
1429
+ "힐": 1425,
1430
+ "힘": 1426,
1431
+ "힙": 1427
1432
+ }