Hugging Face's logo

dl-ru
/
rubert-electra-srl

Token Classification
Transformers
PyTorch
Safetensors
electra
Generated from Trainer
Inference Endpoints
Model card Files Community
dl-ru commited on
Commit
28caf3b
•
1 Parent(s): 79ef8d2

New version with explicit predicate marking

Browse files
Files changed (8) hide show
  1. README.md +111 -0
  2. config.json +77 -0
  3. pytorch_model.bin +3 -0
  4. special_tokens_map.json +7 -0
  5. tokenizer.json +0 -0
  6. tokenizer_config.json +15 -0
  7. training_args.bin +3 -0
  8. vocab.txt +0 -0
README.md ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ base_model: ai-forever/ruElectra-medium
4
+ tags:
5
+ - generated_from_trainer
6
+ model-index:
7
+ - name: rubert-electra-srl
8
+ results: []
9
+ ---
10
+
11
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
12
+ should probably proofread and complete it, then remove this comment. -->
13
+
14
+ # rubert-electra-srl
15
+
16
+ This model is a fine-tuned version of [ai-forever/ruElectra-medium](https://huggingface.co/ai-forever/ruElectra-medium) on an unknown dataset.
17
+ It achieves the following results on the evaluation set:
18
+ - Loss: 0.1501
19
+ - Addressee Precision: 0.9167
20
+ - Addressee Recall: 1.0
21
+ - Addressee F1: 0.9565
22
+ - Addressee Number: 11
23
+ - Benefactive Precision: 1.0
24
+ - Benefactive Recall: 0.5
25
+ - Benefactive F1: 0.6667
26
+ - Benefactive Number: 2
27
+ - Causator Precision: 0.8824
28
+ - Causator Recall: 0.8824
29
+ - Causator F1: 0.8824
30
+ - Causator Number: 17
31
+ - Cause Precision: 0.7778
32
+ - Cause Recall: 0.875
33
+ - Cause F1: 0.8235
34
+ - Cause Number: 8
35
+ - Contrsubject Precision: 0.7273
36
+ - Contrsubject Recall: 0.8421
37
+ - Contrsubject F1: 0.7805
38
+ - Contrsubject Number: 19
39
+ - Deliberative Precision: 0.6667
40
+ - Deliberative Recall: 0.6667
41
+ - Deliberative F1: 0.6667
42
+ - Deliberative Number: 3
43
+ - Destinative Precision: 1.0
44
+ - Destinative Recall: 1.0
45
+ - Destinative F1: 1.0
46
+ - Destinative Number: 1
47
+ - Directivefinal Precision: 0.3333
48
+ - Directivefinal Recall: 0.5
49
+ - Directivefinal F1: 0.4
50
+ - Directivefinal Number: 2
51
+ - Experiencer Precision: 0.8692
52
+ - Experiencer Recall: 0.9208
53
+ - Experiencer F1: 0.8942
54
+ - Experiencer Number: 101
55
+ - Instrument Precision: 1.0
56
+ - Instrument Recall: 0.3333
57
+ - Instrument F1: 0.5
58
+ - Instrument Number: 3
59
+ - Object Precision: 0.8612
60
+ - Object Recall: 0.8866
61
+ - Object F1: 0.8737
62
+ - Object Number: 238
63
+ - Overall Precision: 0.8527
64
+ - Overall Recall: 0.8864
65
+ - Overall F1: 0.8692
66
+ - Overall Accuracy: 0.9711
67
+
68
+ ## Model description
69
+
70
+ More information needed
71
+
72
+ ## Intended uses & limitations
73
+
74
+ More information needed
75
+
76
+ ## Training and evaluation data
77
+
78
+ More information needed
79
+
80
+ ## Training procedure
81
+
82
+ ### Training hyperparameters
83
+
84
+ The following hyperparameters were used during training:
85
+ - learning_rate: 9.81632502988664e-05
86
+ - train_batch_size: 4
87
+ - eval_batch_size: 1
88
+ - seed: 605573
89
+ - gradient_accumulation_steps: 4
90
+ - total_train_batch_size: 16
91
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
92
+ - lr_scheduler_type: linear
93
+ - num_epochs: 5
94
+
95
+ ### Training results
96
+
97
+ | Training Loss | Epoch | Step | Validation Loss | Addressee Precision | Addressee Recall | Addressee F1 | Addressee Number | Benefactive Precision | Benefactive Recall | Benefactive F1 | Benefactive Number | Causator Precision | Causator Recall | Causator F1 | Causator Number | Cause Precision | Cause Recall | Cause F1 | Cause Number | Contrsubject Precision | Contrsubject Recall | Contrsubject F1 | Contrsubject Number | Deliberative Precision | Deliberative Recall | Deliberative F1 | Deliberative Number | Destinative Precision | Destinative Recall | Destinative F1 | Destinative Number | Directivefinal Precision | Directivefinal Recall | Directivefinal F1 | Directivefinal Number | Experiencer Precision | Experiencer Recall | Experiencer F1 | Experiencer Number | Instrument Precision | Instrument Recall | Instrument F1 | Instrument Number | Object Precision | Object Recall | Object F1 | Object Number | Overall Precision | Overall Recall | Overall F1 | Overall Accuracy |
98
+ |:-------------:|:-----:|:----:|:---------------:|:-------------------:|:----------------:|:------------:|:----------------:|:---------------------:|:------------------:|:--------------:|:------------------:|:------------------:|:---------------:|:-----------:|:---------------:|:---------------:|:------------:|:--------:|:------------:|:----------------------:|:-------------------:|:---------------:|:-------------------:|:----------------------:|:-------------------:|:---------------:|:-------------------:|:---------------------:|:------------------:|:--------------:|:------------------:|:------------------------:|:---------------------:|:-----------------:|:---------------------:|:---------------------:|:------------------:|:--------------:|:------------------:|:--------------------:|:-----------------:|:-------------:|:-----------------:|:----------------:|:-------------:|:---------:|:-------------:|:-----------------:|:--------------:|:----------:|:----------------:|
99
+ | 0.1641 | 1.0 | 465 | 0.1697 | 0.9167 | 1.0 | 0.9565 | 11 | 0.0 | 0.0 | 0.0 | 2 | 0.8125 | 0.7647 | 0.7879 | 17 | 0.3333 | 0.375 | 0.3529 | 8 | 0.5926 | 0.8421 | 0.6957 | 19 | 0.0 | 0.0 | 0.0 | 3 | 0.0 | 0.0 | 0.0 | 1 | 0.0 | 0.0 | 0.0 | 2 | 0.8687 | 0.8515 | 0.86 | 101 | 0.0 | 0.0 | 0.0 | 3 | 0.7917 | 0.8782 | 0.8327 | 238 | 0.7916 | 0.8346 | 0.8125 | 0.9619 |
100
+ | 0.0897 | 2.0 | 931 | 0.1417 | 0.9167 | 1.0 | 0.9565 | 11 | 0.0 | 0.0 | 0.0 | 2 | 0.7895 | 0.8824 | 0.8333 | 17 | 0.5714 | 0.5 | 0.5333 | 8 | 0.7083 | 0.8947 | 0.7907 | 19 | 0.6667 | 0.6667 | 0.6667 | 3 | 0.0 | 0.0 | 0.0 | 1 | 0.0 | 0.0 | 0.0 | 2 | 0.8261 | 0.9406 | 0.8796 | 101 | 0.0 | 0.0 | 0.0 | 3 | 0.8275 | 0.8866 | 0.8560 | 238 | 0.8161 | 0.8765 | 0.8452 | 0.9662 |
101
+ | 0.05 | 3.0 | 1396 | 0.1233 | 0.8462 | 1.0 | 0.9167 | 11 | 0.0 | 0.0 | 0.0 | 2 | 0.8333 | 0.8824 | 0.8571 | 17 | 0.5333 | 1.0 | 0.6957 | 8 | 0.8 | 0.8421 | 0.8205 | 19 | 0.5 | 0.6667 | 0.5714 | 3 | 0.0 | 0.0 | 0.0 | 1 | 0.5 | 0.5 | 0.5 | 2 | 0.9565 | 0.8713 | 0.9119 | 101 | 0.0 | 0.0 | 0.0 | 3 | 0.8889 | 0.8739 | 0.8814 | 238 | 0.8769 | 0.8617 | 0.8692 | 0.9728 |
102
+ | 0.0381 | 4.0 | 1862 | 0.1380 | 0.8462 | 1.0 | 0.9167 | 11 | 0.0 | 0.0 | 0.0 | 2 | 0.8824 | 0.8824 | 0.8824 | 17 | 0.7778 | 0.875 | 0.8235 | 8 | 0.7727 | 0.8947 | 0.8293 | 19 | 0.4 | 0.6667 | 0.5 | 3 | 0.0 | 0.0 | 0.0 | 1 | 0.3333 | 0.5 | 0.4 | 2 | 0.8980 | 0.8713 | 0.8844 | 101 | 0.0 | 0.0 | 0.0 | 3 | 0.8814 | 0.8739 | 0.8776 | 238 | 0.8660 | 0.8617 | 0.8639 | 0.9711 |
103
+ | 0.0271 | 4.99 | 2325 | 0.1501 | 0.9167 | 1.0 | 0.9565 | 11 | 1.0 | 0.5 | 0.6667 | 2 | 0.8824 | 0.8824 | 0.8824 | 17 | 0.7778 | 0.875 | 0.8235 | 8 | 0.7273 | 0.8421 | 0.7805 | 19 | 0.6667 | 0.6667 | 0.6667 | 3 | 1.0 | 1.0 | 1.0 | 1 | 0.3333 | 0.5 | 0.4 | 2 | 0.8692 | 0.9208 | 0.8942 | 101 | 1.0 | 0.3333 | 0.5 | 3 | 0.8612 | 0.8866 | 0.8737 | 238 | 0.8527 | 0.8864 | 0.8692 | 0.9711 |
104
+
105
+
106
+ ### Framework versions
107
+
108
+ - Transformers 4.33.2
109
+ - Pytorch 2.0.1+cu117
110
+ - Datasets 2.14.5
111
+ - Tokenizers 0.13.3
config.json ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "ai-forever/ruElectra-medium",
3
+ "architectures": [
4
+ "ElectraForTokenClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "embedding_size": 576,
9
+ "generator_size": "0.25",
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 576,
13
+ "id2label": {
14
+ "0": "O",
15
+ "1": "B-Object",
16
+ "2": "B-Experiencer",
17
+ "3": "B-Cause",
18
+ "4": "B-Deliberative",
19
+ "5": "B-Causator",
20
+ "6": "B-ContrSubject",
21
+ "7": "B-Benefactive",
22
+ "8": "B-Addressee",
23
+ "9": "I-Object",
24
+ "10": "B-Destinative",
25
+ "11": "I-ContrSubject",
26
+ "12": "B-Instrument",
27
+ "13": "I-Deliberative",
28
+ "14": "B-DirectiveFinal",
29
+ "15": "B-Mediative",
30
+ "16": "I-DirectiveFinal",
31
+ "17": "B-DirectiveInitial",
32
+ "18": "I-DirectiveInitial",
33
+ "19": "I-Experiencer",
34
+ "20": "I-Cause"
35
+ },
36
+ "initializer_range": 0.02,
37
+ "intermediate_size": 2304,
38
+ "label2id": {
39
+ "B-Addressee": 8,
40
+ "B-Benefactive": 7,
41
+ "B-Causator": 5,
42
+ "B-Cause": 3,
43
+ "B-ContrSubject": 6,
44
+ "B-Deliberative": 4,
45
+ "B-Destinative": 10,
46
+ "B-DirectiveFinal": 14,
47
+ "B-DirectiveInitial": 17,
48
+ "B-Experiencer": 2,
49
+ "B-Instrument": 12,
50
+ "B-Mediative": 15,
51
+ "B-Object": 1,
52
+ "I-Cause": 20,
53
+ "I-ContrSubject": 11,
54
+ "I-Deliberative": 13,
55
+ "I-DirectiveFinal": 16,
56
+ "I-DirectiveInitial": 18,
57
+ "I-Experiencer": 19,
58
+ "I-Object": 9,
59
+ "O": 0
60
+ },
61
+ "layer_norm_eps": 1e-12,
62
+ "max_position_embeddings": 512,
63
+ "model_type": "electra",
64
+ "num_attention_heads": 9,
65
+ "num_hidden_layers": 12,
66
+ "pad_token_id": 0,
67
+ "position_embedding_type": "absolute",
68
+ "summary_activation": "gelu",
69
+ "summary_last_dropout": 0.1,
70
+ "summary_type": "first",
71
+ "summary_use_proj": true,
72
+ "torch_dtype": "float32",
73
+ "transformers_version": "4.33.2",
74
+ "type_vocab_size": 2,
75
+ "use_cache": true,
76
+ "vocab_size": 64000
77
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76969ff90e59d63a283f29d79849d50a5e46a1fb08fcd19317d3946f70d38c5f
3
+ size 340224041
special_tokens_map.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": "[CLS]",
3
+ "mask_token": "[MASK]",
4
+ "pad_token": "[PAD]",
5
+ "sep_token": "[SEP]",
6
+ "unk_token": "[UNK]"
7
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "clean_up_tokenization_spaces": true,
3
+ "cls_token": "[CLS]",
4
+ "do_basic_tokenize": true,
5
+ "do_lower_case": true,
6
+ "mask_token": "[MASK]",
7
+ "model_max_length": 1000000000000000019884624838656,
8
+ "never_split": null,
9
+ "pad_token": "[PAD]",
10
+ "sep_token": "[SEP]",
11
+ "strip_accents": null,
12
+ "tokenize_chinese_chars": true,
13
+ "tokenizer_class": "ElectraTokenizer",
14
+ "unk_token": "[UNK]"
15
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:75edf4b7e1d1ebc4ebe4474ec469d7a7b5fb8850237ebcdc4d5bfc3137a6e8c8
3
+ size 4155
vocab.txt ADDED
The diff for this file is too large to render. See raw diff