Hugging Face's logo

BUCOLIN
/
HistBERTurk-dependency-parsing

PyTorch
Model card Files Community
1
beratdogan commited on
Commit
a12b352
·
verified ·
1 Parent(s): 19e5a92

Upload folder using huggingface_hub

Browse files
Files changed (8) hide show
  1. config.json +128 -0
  2. info.log +0 -0
  3. labels.vocab +60 -0
  4. pytorch_model.bin +3 -0
  5. tokenizer/special_tokens_map.json +1 -0
  6. tokenizer/tokenizer_config.json +1 -0
  7. tokenizer/vocab.txt +0 -0
  8. transformer.json +21 -0
config.json ADDED
@@ -0,0 +1,128 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "experiment": "ota_boun",
3
+ "name": "bert-base-turkish-cased",
4
+ "n_gpu": 1,
5
+ "model": {
6
+ "type": "MultiParser",
7
+ "args": {
8
+ "embeddings_processor": {
9
+ "type": "BertWrapper",
10
+ "args": {
11
+ "model_path": "data/pretrained_embeddings/bert-base-turkish-cased-nonpretrained",
12
+ "fine_tune": true,
13
+ "hidden_dropout": 0.2,
14
+ "attn_dropout": 0.2,
15
+ "output_dropout": 0.5,
16
+ "scalar_mix_layer_dropout": 0.1,
17
+ "token_mask_prob": 0.15
18
+ }
19
+ },
20
+ "outputs": {
21
+ "heads": {
22
+ "type": "ArcScorer",
23
+ "args": {
24
+ "scorer_class": "DeepBiaffineScorer",
25
+ "head_mode": "single_head",
26
+ "hidden_size": 768,
27
+ "dropout": 0.33,
28
+ "vocab": {
29
+ "type": "IntegerVocab"
30
+ }
31
+ }
32
+ },
33
+ "labels": {
34
+ "type": "DependencyClassifier",
35
+ "args": {
36
+ "scorer_class": "DeepBiaffineScorer",
37
+ "hidden_size": 256,
38
+ "dropout": 0.33,
39
+ "vocab": {
40
+ "type": "BasicVocab",
41
+ "args": {
42
+ "vocab_filename": "data/corpora/ota_boun/vocab/basic.vocab"
43
+ }
44
+ }
45
+ }
46
+ }
47
+ },
48
+ "post_processors": [
49
+ {
50
+ "type": "FactorizedMSTPostProcessor",
51
+ "args": {
52
+ "annotation_ids": [
53
+ "heads",
54
+ "labels"
55
+ ]
56
+ }
57
+ }
58
+ ]
59
+ }
60
+ },
61
+ "data_loaders": {
62
+ "type": "BucketedCONLLLoader",
63
+ "args": {
64
+ "annotation_layers": {
65
+ "heads": {
66
+ "type": "TagSequence",
67
+ "source_column": 6,
68
+ "args": {
69
+ "ignore_root": true
70
+ }
71
+ },
72
+ "labels": {
73
+ "type": "DependencyMatrix",
74
+ "source_column": [
75
+ 6,
76
+ 7
77
+ ],
78
+ "args": {
79
+ "ignore_non_relations": true
80
+ }
81
+ }
82
+ },
83
+ "batch_size": 32,
84
+ "bucket_size": 256,
85
+ "max_tokens_per_batch": 156800,
86
+ "num_workers": 2
87
+ },
88
+ "paths": {
89
+ "train": "data/corpora/ota_boun/ota_boun-ud-train.conllu",
90
+ "dev": "data/corpora/ota_boun/ota_boun-ud-test.conllu",
91
+ "test": "data/corpora/ota_boun/tr_boun-ud-test.conllu"
92
+ }
93
+ },
94
+ "trainer": {
95
+ "min_epochs": 15,
96
+ "max_epochs": 300,
97
+ "early_stop": 15,
98
+ "save_dir": "data/saved_models/",
99
+ "save_period": 20,
100
+ "verbosity": 2,
101
+ "validation_criterion": {
102
+ "metrics": {
103
+ "heads": "fscore",
104
+ "labels": "fscore"
105
+ },
106
+ "weighting": "multiplicative"
107
+ },
108
+ "optimizer": {
109
+ "type": "AdamW",
110
+ "args": {
111
+ "lr": 4e-05,
112
+ "weight_decay": 0.0
113
+ }
114
+ },
115
+ "lr_scheduler": {
116
+ "type": "LambdaLR",
117
+ "args": {
118
+ "lr_lambda": "SqrtSchedule(400)"
119
+ }
120
+ },
121
+ "loss": {
122
+ "type": "CrossEntropyLoss",
123
+ "args": {
124
+ "ignore_index": -1
125
+ }
126
+ }
127
+ }
128
+ }
info.log ADDED
The diff for this file is too large to render. See raw diff
 
labels.vocab ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ root
2
+ case
3
+ det
4
+ obl
5
+ nsubj
6
+ punct
7
+ flat
8
+ obj
9
+ advcl
10
+ nummod
11
+ mark
12
+ amod
13
+ nmod
14
+ compound
15
+ advmod
16
+ nsubj:pass
17
+ aux
18
+ aux:pass
19
+ acl
20
+ conj
21
+ obl:tmod
22
+ cc
23
+ det:predet
24
+ cop
25
+ nmod:poss
26
+ ccomp
27
+ parataxis
28
+ expl
29
+ flat:foreign
30
+ xcomp
31
+ csubj
32
+ fixed
33
+ acl:relcl
34
+ appos
35
+ compound:prt
36
+ iobj
37
+ nmod:tmod
38
+ obl:npmod
39
+ cc:preconj
40
+ reparandum
41
+ vocative
42
+ discourse
43
+ goeswith
44
+ list
45
+ nmod:npmod
46
+ orphan
47
+ dep
48
+ csubj:pass
49
+ dislocated
50
+ compound:lvc
51
+ advmod:emph
52
+ obl:agent
53
+ compound:redup
54
+ aux:q
55
+ clf
56
+ csubj:outer
57
+ dep:der
58
+ discourse:q
59
+ nmod:part
60
+ nsubj:outer
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e1c1cd47eff0dc4b0698f0835b48c99415147495373fca4d27fc030e41c5c19
3
+ size 1396526621
tokenizer/special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"do_lower_case": false, "max_len": 512, "special_tokens_map_file": null, "full_tokenizer_file": null}
tokenizer/vocab.txt ADDED
The diff for this file is too large to render. See raw diff
 
transformer.json ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "BertModel"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.2,
6
+ "gradient_checkpointing": false,
7
+ "hidden_act": "gelu",
8
+ "hidden_dropout_prob": 0.2,
9
+ "hidden_size": 768,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 3072,
12
+ "layer_norm_eps": 1e-12,
13
+ "max_position_embeddings": 512,
14
+ "model_type": "bert",
15
+ "num_attention_heads": 12,
16
+ "num_hidden_layers": 12,
17
+ "output_hidden_states": true,
18
+ "pad_token_id": 0,
19
+ "type_vocab_size": 2,
20
+ "vocab_size": 32000
21
+ }