Hugging Face's logo

rcds
/
MiniLM-swiss_citation_extraction-de-fr-it

Token Classification
Transformers
PyTorch
Safetensors
German
French
Italian
bert
Inference Endpoints
Model card Files Community
2

First commit

#1
by kapllan - opened
base: refs/heads/main
←
from: refs/pr/1
Discussion Files changed
+1042
-0
Files changed (14) hide show
  1. .gitattributes +3 -0
  2. all_results.json +105 -0
  3. config.json +41 -0
  4. predict_results.json +83 -0
  5. pytorch_model.bin +3 -0
  6. sentencepiece.bpe.model +3 -0
  7. special_tokens_map.json +9 -0
  8. test_predictions_detailed.csv +3 -0
  9. test_predictions_detailed.json +3 -0
  10. tokenizer.json +3 -0
  11. tokenizer_config.json +20 -0
  12. train_results.json +24 -0
  13. trainer_state.json +739 -0
  14. training_args.bin +3 -0
.gitattributes CHANGED
@@ -32,3 +32,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
32
  *.zip filter=lfs diff=lfs merge=lfs -text
33
  *.zst filter=lfs diff=lfs merge=lfs -text
34
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
32
  *.zip filter=lfs diff=lfs merge=lfs -text
33
  *.zst filter=lfs diff=lfs merge=lfs -text
34
  *tfevents* filter=lfs diff=lfs merge=lfs -text
35
+ test_predictions_detailed.csv filter=lfs diff=lfs merge=lfs -text
36
+ test_predictions_detailed.json filter=lfs diff=lfs merge=lfs -text
37
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
all_results.json ADDED
@@ -0,0 +1,105 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "de_predict/_CITATION_f1": 0.9793131792857794,
3
+ "de_predict/_CITATION_precision": 0.9852522282458881,
4
+ "de_predict/_CITATION_recall": 0.9734453018610985,
5
+ "de_predict/_LAW_f1": 0.9207842961099632,
6
+ "de_predict/_LAW_precision": 0.8598544432559407,
7
+ "de_predict/_LAW_recall": 0.9910077594333921,
8
+ "de_predict/_accuracy_normalized": 0.9880353522464387,
9
+ "de_predict/_loss": 0.035749662667512894,
10
+ "de_predict/_macro-f1": 0.9504272924171073,
11
+ "de_predict/_macro-precision": 0.9822265306472453,
12
+ "de_predict/_macro-recall": 0.9232171398568052,
13
+ "de_predict/_micro-f1": 0.9405898834091524,
14
+ "de_predict/_micro-precision": 0.9849051246865093,
15
+ "de_predict/_micro-recall": 0.9000908134288556,
16
+ "de_predict/_runtime": 265.8036,
17
+ "de_predict/_samples_per_second": 70.277,
18
+ "de_predict/_steps_per_second": 0.549,
19
+ "de_predict/_weighted-f1": 0.939658320951984,
20
+ "de_predict/_weighted-precision": 0.9854977355183103,
21
+ "de_predict/_weighted-recall": 0.9000908134288556,
22
+ "epoch": 23.0,
23
+ "fr_predict/_CITATION_f1": 0.9554686901203342,
24
+ "fr_predict/_CITATION_precision": 0.9684586699813549,
25
+ "fr_predict/_CITATION_recall": 0.9428225684465286,
26
+ "fr_predict/_LAW_f1": 0.910095519316377,
27
+ "fr_predict/_LAW_precision": 0.8366717393986756,
28
+ "fr_predict/_LAW_recall": 0.9976459048553212,
29
+ "fr_predict/_accuracy_normalized": 0.9830767480044869,
30
+ "fr_predict/_loss": 0.04828711599111557,
31
+ "fr_predict/_macro-f1": 0.9330080903677362,
32
+ "fr_predict/_macro-precision": 0.9702342366509249,
33
+ "fr_predict/_macro-recall": 0.9029739799827206,
34
+ "fr_predict/_micro-f1": 0.920617324580396,
35
+ "fr_predict/_micro-precision": 0.9842228065627199,
36
+ "fr_predict/_micro-recall": 0.8647338279317974,
37
+ "fr_predict/_runtime": 102.8006,
38
+ "fr_predict/_samples_per_second": 74.863,
39
+ "fr_predict/_steps_per_second": 0.593,
40
+ "fr_predict/_weighted-f1": 0.9198669665372888,
41
+ "fr_predict/_weighted-precision": 0.9861681830521788,
42
+ "fr_predict/_weighted-recall": 0.8647338279317974,
43
+ "it_predict/_CITATION_f1": 0.9703896103896105,
44
+ "it_predict/_CITATION_precision": 0.9769874476987448,
45
+ "it_predict/_CITATION_recall": 0.9638802889576883,
46
+ "it_predict/_LAW_f1": 0.9099276791584483,
47
+ "it_predict/_LAW_precision": 0.8422590068159689,
48
+ "it_predict/_LAW_recall": 0.9894195024306548,
49
+ "it_predict/_accuracy_normalized": 0.9892137683075134,
50
+ "it_predict/_loss": 0.031956374645233154,
51
+ "it_predict/_macro-f1": 0.9413484848298093,
52
+ "it_predict/_macro-precision": 0.9766498956941716,
53
+ "it_predict/_macro-recall": 0.9119834901073706,
54
+ "it_predict/_micro-f1": 0.9311429570080392,
55
+ "it_predict/_micro-precision": 0.9803127874885005,
56
+ "it_predict/_micro-recall": 0.8866699950074888,
57
+ "it_predict/_runtime": 26.6394,
58
+ "it_predict/_samples_per_second": 69.146,
59
+ "it_predict/_steps_per_second": 0.563,
60
+ "it_predict/_weighted-f1": 0.929971077318579,
61
+ "it_predict/_weighted-precision": 0.9813271971464931,
62
+ "it_predict/_weighted-recall": 0.8866699950074888,
63
+ "predict/_CITATION_f1": 0.973621340187501,
64
+ "predict/_CITATION_precision": 0.981138340970977,
65
+ "predict/_CITATION_recall": 0.9662186467837405,
66
+ "predict/_LAW_f1": 0.9168199439712499,
67
+ "predict/_LAW_precision": 0.8514980289093298,
68
+ "predict/_LAW_recall": 0.9929968125536349,
69
+ "predict/_accuracy_normalized": 0.986841752305624,
70
+ "predict/_loss": 0.03871648758649826,
71
+ "predict/_macro-f1": 0.9455976917351873,
72
+ "predict/_macro-precision": 0.9796077296686877,
73
+ "predict/_macro-recall": 0.9169959471957758,
74
+ "predict/_micro-f1": 0.934344809828224,
75
+ "predict/_micro-precision": 0.9844524443053164,
76
+ "predict/_micro-recall": 0.8890909776278342,
77
+ "predict/_runtime": 396.5316,
78
+ "predict/_samples_per_second": 71.162,
79
+ "predict/_steps_per_second": 0.557,
80
+ "predict/_weighted-f1": 0.9333974918752409,
81
+ "predict/_weighted-precision": 0.9854002360022739,
82
+ "predict/_weighted-recall": 0.8890909776278342,
83
+ "predict_samples": 28218,
84
+ "train_CITATION_f1": 0.9814787122684838,
85
+ "train_CITATION_precision": 0.9844650624611626,
86
+ "train_CITATION_recall": 0.978510425318936,
87
+ "train_LAW_f1": 0.9211885809936669,
88
+ "train_LAW_precision": 0.8550646934923544,
89
+ "train_LAW_recall": 0.9983966169255928,
90
+ "train_accuracy_normalized": 0.9905667906095162,
91
+ "train_loss": 0.026775669306516647,
92
+ "train_macro-f1": 0.9515974221562933,
93
+ "train_macro-precision": 0.9884535211222644,
94
+ "train_macro-recall": 0.9202279237061262,
95
+ "train_micro-f1": 0.9429566347321088,
96
+ "train_micro-precision": 0.9909403074465961,
97
+ "train_micro-recall": 0.8994053033460294,
98
+ "train_runtime": 1297.076,
99
+ "train_samples": 90101,
100
+ "train_samples_per_second": 69.465,
101
+ "train_steps_per_second": 0.543,
102
+ "train_weighted-f1": 0.9419101385291693,
103
+ "train_weighted-precision": 0.9916707876514618,
104
+ "train_weighted-recall": 0.8994053033460294
105
+ }
config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "microsoft/Multilingual-MiniLM-L12-H384",
3
+ "architectures": [
4
+ "BertForTokenClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
8
+ "finetuning_task": "all_original",
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 384,
12
+ "id2label": {
13
+ "0": "0",
14
+ "1": "B-CITATION",
15
+ "2": "I-CITATION",
16
+ "3": "B-LAW",
17
+ "4": "I-LAW"
18
+ },
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 1536,
21
+ "label2id": {
22
+ "LABEL_0": "O",
23
+ "LABEL_1": "B-CITATION",
24
+ "LABEL_2": "I-CITATION",
25
+ "LABEL_3": "B-LAW",
26
+ "LABEL_4": "I-LAW"
27
+ },
28
+ "layer_norm_eps": 1e-12,
29
+ "max_position_embeddings": 512,
30
+ "model_type": "bert",
31
+ "num_attention_heads": 12,
32
+ "num_hidden_layers": 12,
33
+ "pad_token_id": 0,
34
+ "position_embedding_type": "absolute",
35
+ "tokenizer_class": "XLMRobertaTokenizer",
36
+ "torch_dtype": "float16",
37
+ "transformers_version": "4.28.1",
38
+ "type_vocab_size": 2,
39
+ "use_cache": true,
40
+ "vocab_size": 250037
41
+ }
predict_results.json ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "de_predict/_CITATION_f1": 0.9793131792857794,
3
+ "de_predict/_CITATION_precision": 0.9852522282458881,
4
+ "de_predict/_CITATION_recall": 0.9734453018610985,
5
+ "de_predict/_LAW_f1": 0.9207842961099632,
6
+ "de_predict/_LAW_precision": 0.8598544432559407,
7
+ "de_predict/_LAW_recall": 0.9910077594333921,
8
+ "de_predict/_accuracy_normalized": 0.9880353522464387,
9
+ "de_predict/_loss": 0.035749662667512894,
10
+ "de_predict/_macro-f1": 0.9504272924171073,
11
+ "de_predict/_macro-precision": 0.9822265306472453,
12
+ "de_predict/_macro-recall": 0.9232171398568052,
13
+ "de_predict/_micro-f1": 0.9405898834091524,
14
+ "de_predict/_micro-precision": 0.9849051246865093,
15
+ "de_predict/_micro-recall": 0.9000908134288556,
16
+ "de_predict/_runtime": 265.8036,
17
+ "de_predict/_samples_per_second": 70.277,
18
+ "de_predict/_steps_per_second": 0.549,
19
+ "de_predict/_weighted-f1": 0.939658320951984,
20
+ "de_predict/_weighted-precision": 0.9854977355183103,
21
+ "de_predict/_weighted-recall": 0.9000908134288556,
22
+ "fr_predict/_CITATION_f1": 0.9554686901203342,
23
+ "fr_predict/_CITATION_precision": 0.9684586699813549,
24
+ "fr_predict/_CITATION_recall": 0.9428225684465286,
25
+ "fr_predict/_LAW_f1": 0.910095519316377,
26
+ "fr_predict/_LAW_precision": 0.8366717393986756,
27
+ "fr_predict/_LAW_recall": 0.9976459048553212,
28
+ "fr_predict/_accuracy_normalized": 0.9830767480044869,
29
+ "fr_predict/_loss": 0.04828711599111557,
30
+ "fr_predict/_macro-f1": 0.9330080903677362,
31
+ "fr_predict/_macro-precision": 0.9702342366509249,
32
+ "fr_predict/_macro-recall": 0.9029739799827206,
33
+ "fr_predict/_micro-f1": 0.920617324580396,
34
+ "fr_predict/_micro-precision": 0.9842228065627199,
35
+ "fr_predict/_micro-recall": 0.8647338279317974,
36
+ "fr_predict/_runtime": 102.8006,
37
+ "fr_predict/_samples_per_second": 74.863,
38
+ "fr_predict/_steps_per_second": 0.593,
39
+ "fr_predict/_weighted-f1": 0.9198669665372888,
40
+ "fr_predict/_weighted-precision": 0.9861681830521788,
41
+ "fr_predict/_weighted-recall": 0.8647338279317974,
42
+ "it_predict/_CITATION_f1": 0.9703896103896105,
43
+ "it_predict/_CITATION_precision": 0.9769874476987448,
44
+ "it_predict/_CITATION_recall": 0.9638802889576883,
45
+ "it_predict/_LAW_f1": 0.9099276791584483,
46
+ "it_predict/_LAW_precision": 0.8422590068159689,
47
+ "it_predict/_LAW_recall": 0.9894195024306548,
48
+ "it_predict/_accuracy_normalized": 0.9892137683075134,
49
+ "it_predict/_loss": 0.031956374645233154,
50
+ "it_predict/_macro-f1": 0.9413484848298093,
51
+ "it_predict/_macro-precision": 0.9766498956941716,
52
+ "it_predict/_macro-recall": 0.9119834901073706,
53
+ "it_predict/_micro-f1": 0.9311429570080392,
54
+ "it_predict/_micro-precision": 0.9803127874885005,
55
+ "it_predict/_micro-recall": 0.8866699950074888,
56
+ "it_predict/_runtime": 26.6394,
57
+ "it_predict/_samples_per_second": 69.146,
58
+ "it_predict/_steps_per_second": 0.563,
59
+ "it_predict/_weighted-f1": 0.929971077318579,
60
+ "it_predict/_weighted-precision": 0.9813271971464931,
61
+ "it_predict/_weighted-recall": 0.8866699950074888,
62
+ "predict/_CITATION_f1": 0.973621340187501,
63
+ "predict/_CITATION_precision": 0.981138340970977,
64
+ "predict/_CITATION_recall": 0.9662186467837405,
65
+ "predict/_LAW_f1": 0.9168199439712499,
66
+ "predict/_LAW_precision": 0.8514980289093298,
67
+ "predict/_LAW_recall": 0.9929968125536349,
68
+ "predict/_accuracy_normalized": 0.986841752305624,
69
+ "predict/_loss": 0.03871648758649826,
70
+ "predict/_macro-f1": 0.9455976917351873,
71
+ "predict/_macro-precision": 0.9796077296686877,
72
+ "predict/_macro-recall": 0.9169959471957758,
73
+ "predict/_micro-f1": 0.934344809828224,
74
+ "predict/_micro-precision": 0.9844524443053164,
75
+ "predict/_micro-recall": 0.8890909776278342,
76
+ "predict/_runtime": 396.5316,
77
+ "predict/_samples_per_second": 71.162,
78
+ "predict/_steps_per_second": 0.557,
79
+ "predict/_weighted-f1": 0.9333974918752409,
80
+ "predict/_weighted-precision": 0.9854002360022739,
81
+ "predict/_weighted-recall": 0.8890909776278342,
82
+ "predict_samples": 28218
83
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db557e55138656512eaf5ccb97ae3df00697588e8157f005996d982a963fe9db
3
+ size 235087853
sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
3
+ size 5069051
special_tokens_map.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": "<mask>",
6
+ "pad_token": "<pad>",
7
+ "sep_token": "</s>",
8
+ "unk_token": "<unk>"
9
+ }
test_predictions_detailed.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5189f91e0ed539a87b24b370cbe66c7f00336dc4d4a734cc66b2cca812580d1
3
+ size 138154683
test_predictions_detailed.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98d732e378ec11d5346171b8e9b1d5e9e266b77ec17192d7e366ad128f1f71a6
3
+ size 960496838
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2bae0289b9bb66bfbf85d834f6313ebd8e40f484df2bd0351539e74a0e4ddeb4
3
+ size 17098081
tokenizer_config.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "clean_up_tokenization_spaces": true,
4
+ "cls_token": "<s>",
5
+ "eos_token": "</s>",
6
+ "mask_token": {
7
+ "__type": "AddedToken",
8
+ "content": "<mask>",
9
+ "lstrip": true,
10
+ "normalized": true,
11
+ "rstrip": false,
12
+ "single_word": false
13
+ },
14
+ "model_max_length": 1000000000000000019884624838656,
15
+ "pad_token": "<pad>",
16
+ "sep_token": "</s>",
17
+ "sp_model_kwargs": {},
18
+ "tokenizer_class": "XLMRobertaTokenizer",
19
+ "unk_token": "<unk>"
20
+ }
train_results.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 23.0,
3
+ "train_CITATION_f1": 0.9814787122684838,
4
+ "train_CITATION_precision": 0.9844650624611626,
5
+ "train_CITATION_recall": 0.978510425318936,
6
+ "train_LAW_f1": 0.9211885809936669,
7
+ "train_LAW_precision": 0.8550646934923544,
8
+ "train_LAW_recall": 0.9983966169255928,
9
+ "train_accuracy_normalized": 0.9905667906095162,
10
+ "train_loss": 0.026775669306516647,
11
+ "train_macro-f1": 0.9515974221562933,
12
+ "train_macro-precision": 0.9884535211222644,
13
+ "train_macro-recall": 0.9202279237061262,
14
+ "train_micro-f1": 0.9429566347321088,
15
+ "train_micro-precision": 0.9909403074465961,
16
+ "train_micro-recall": 0.8994053033460294,
17
+ "train_runtime": 1297.076,
18
+ "train_samples": 90101,
19
+ "train_samples_per_second": 69.465,
20
+ "train_steps_per_second": 0.543,
21
+ "train_weighted-f1": 0.9419101385291693,
22
+ "train_weighted-precision": 0.9916707876514618,
23
+ "train_weighted-recall": 0.8994053033460294
24
+ }
trainer_state.json ADDED
@@ -0,0 +1,739 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.038164205849170685,
3
+ "best_model_checkpoint": "swiss_citation_extraction/original/microsoft/Multilingual-MiniLM-L12-H384/seed_2/checkpoint-12672",
4
+ "epoch": 23.0,
5
+ "global_step": 16192,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 1.0,
12
+ "learning_rate": 9.800000000000001e-06,
13
+ "loss": 0.4621,
14
+ "step": 704
15
+ },
16
+ {
17
+ "epoch": 1.0,
18
+ "eval_CITATION_f1": 0.9656679647142971,
19
+ "eval_CITATION_precision": 0.9683614918712146,
20
+ "eval_CITATION_recall": 0.9629893802504359,
21
+ "eval_LAW_f1": 0.8924282681945703,
22
+ "eval_LAW_precision": 0.8161431325170364,
23
+ "eval_LAW_recall": 0.9844445899921405,
24
+ "eval_accuracy_normalized": 0.9826788171520043,
25
+ "eval_loss": 0.22096700966358185,
26
+ "eval_macro-f1": 0.9328819776713801,
27
+ "eval_macro-precision": 0.9737169851212881,
28
+ "eval_macro-recall": 0.8988391632624153,
29
+ "eval_micro-f1": 0.9176521739130435,
30
+ "eval_micro-precision": 0.9781712008156833,
31
+ "eval_micro-recall": 0.8641853990091307,
32
+ "eval_runtime": 183.4312,
33
+ "eval_samples_per_second": 69.427,
34
+ "eval_steps_per_second": 0.545,
35
+ "eval_weighted-f1": 0.9165196950110328,
36
+ "eval_weighted-precision": 0.9789422067379471,
37
+ "eval_weighted-recall": 0.8641853990091307,
38
+ "step": 704
39
+ },
40
+ {
41
+ "epoch": 2.0,
42
+ "learning_rate": 9.600000000000001e-06,
43
+ "loss": 0.1514,
44
+ "step": 1408
45
+ },
46
+ {
47
+ "epoch": 2.0,
48
+ "eval_CITATION_f1": 0.967408173823073,
49
+ "eval_CITATION_precision": 0.9715450403644793,
50
+ "eval_CITATION_recall": 0.9633063877001109,
51
+ "eval_LAW_f1": 0.9143357699510664,
52
+ "eval_LAW_precision": 0.8484695593676421,
53
+ "eval_LAW_recall": 0.9912889703955986,
54
+ "eval_accuracy_normalized": 0.9859679337789155,
55
+ "eval_loss": 0.12093725800514221,
56
+ "eval_macro-f1": 0.9415014323169688,
57
+ "eval_macro-precision": 0.9772976790478547,
58
+ "eval_macro-recall": 0.9111432819443076,
59
+ "eval_micro-f1": 0.9296999978086035,
60
+ "eval_micro-precision": 0.9831070120962135,
61
+ "eval_micro-recall": 0.881796641170602,
62
+ "eval_runtime": 182.5229,
63
+ "eval_samples_per_second": 69.772,
64
+ "eval_steps_per_second": 0.548,
65
+ "eval_weighted-f1": 0.9288920564740497,
66
+ "eval_weighted-precision": 0.9840176399820014,
67
+ "eval_weighted-recall": 0.881796641170602,
68
+ "step": 1408
69
+ },
70
+ {
71
+ "epoch": 3.0,
72
+ "learning_rate": 9.4e-06,
73
+ "loss": 0.0915,
74
+ "step": 2112
75
+ },
76
+ {
77
+ "epoch": 3.0,
78
+ "eval_CITATION_f1": 0.967600700525394,
79
+ "eval_CITATION_precision": 0.9719334719334719,
80
+ "eval_CITATION_recall": 0.9633063877001109,
81
+ "eval_LAW_f1": 0.9151176825588413,
82
+ "eval_LAW_precision": 0.8484500895255148,
83
+ "eval_LAW_recall": 0.9931556195965417,
84
+ "eval_accuracy_normalized": 0.9861328466436042,
85
+ "eval_loss": 0.08684782683849335,
86
+ "eval_macro-f1": 0.941871559216277,
87
+ "eval_macro-precision": 0.9782310036483264,
88
+ "eval_macro-recall": 0.9110962962834105,
89
+ "eval_micro-f1": 0.9302254289061389,
90
+ "eval_micro-precision": 0.9844278630022709,
91
+ "eval_micro-recall": 0.8816802606728514,
92
+ "eval_runtime": 182.8851,
93
+ "eval_samples_per_second": 69.634,
94
+ "eval_steps_per_second": 0.547,
95
+ "eval_weighted-f1": 0.9294201322268216,
96
+ "eval_weighted-precision": 0.985410666040948,
97
+ "eval_weighted-recall": 0.8816802606728514,
98
+ "step": 2112
99
+ },
100
+ {
101
+ "epoch": 4.0,
102
+ "learning_rate": 9.200000000000002e-06,
103
+ "loss": 0.0666,
104
+ "step": 2816
105
+ },
106
+ {
107
+ "epoch": 4.0,
108
+ "eval_CITATION_f1": 0.9689891581632653,
109
+ "eval_CITATION_precision": 0.9747393744987971,
110
+ "eval_CITATION_recall": 0.9633063877001109,
111
+ "eval_LAW_f1": 0.9185858127034597,
112
+ "eval_LAW_precision": 0.854262299698668,
113
+ "eval_LAW_recall": 0.9933848572177102,
114
+ "eval_accuracy_normalized": 0.986596354625882,
115
+ "eval_loss": 0.06908567994832993,
116
+ "eval_macro-f1": 0.9441561554169013,
117
+ "eval_macro-precision": 0.9783456224589105,
118
+ "eval_macro-recall": 0.9151899141069411,
119
+ "eval_micro-f1": 0.9328400807938877,
120
+ "eval_micro-precision": 0.984590072762664,
121
+ "eval_micro-recall": 0.8862583955613033,
122
+ "eval_runtime": 182.9721,
123
+ "eval_samples_per_second": 69.601,
124
+ "eval_steps_per_second": 0.547,
125
+ "eval_weighted-f1": 0.9320652823900626,
126
+ "eval_weighted-precision": 0.9855669145232293,
127
+ "eval_weighted-recall": 0.8862583955613033,
128
+ "step": 2816
129
+ },
130
+ {
131
+ "epoch": 5.0,
132
+ "learning_rate": 9e-06,
133
+ "loss": 0.0536,
134
+ "step": 3520
135
+ },
136
+ {
137
+ "epoch": 5.0,
138
+ "eval_CITATION_f1": 0.9693755482893373,
139
+ "eval_CITATION_precision": 0.9755216693418941,
140
+ "eval_CITATION_recall": 0.9633063877001109,
141
+ "eval_LAW_f1": 0.9195075040511275,
142
+ "eval_LAW_precision": 0.8552753909001268,
143
+ "eval_LAW_recall": 0.9941708147760021,
144
+ "eval_accuracy_normalized": 0.9867220387952753,
145
+ "eval_loss": 0.059190474450588226,
146
+ "eval_macro-f1": 0.9446558923319595,
147
+ "eval_macro-precision": 0.9787386012380566,
148
+ "eval_macro-recall": 0.9157752849313606,
149
+ "eval_micro-f1": 0.9335309617918313,
150
+ "eval_micro-precision": 0.9851462205125828,
151
+ "eval_micro-recall": 0.887055043191587,
152
+ "eval_runtime": 182.5981,
153
+ "eval_samples_per_second": 69.743,
154
+ "eval_steps_per_second": 0.548,
155
+ "eval_weighted-f1": 0.932770084197937,
156
+ "eval_weighted-precision": 0.9861471971793212,
157
+ "eval_weighted-recall": 0.887055043191587,
158
+ "step": 3520
159
+ },
160
+ {
161
+ "epoch": 6.0,
162
+ "learning_rate": 8.8e-06,
163
+ "loss": 0.0458,
164
+ "step": 4224
165
+ },
166
+ {
167
+ "epoch": 6.0,
168
+ "eval_CITATION_f1": 0.9689891581632653,
169
+ "eval_CITATION_precision": 0.9747393744987971,
170
+ "eval_CITATION_recall": 0.9633063877001109,
171
+ "eval_LAW_f1": 0.9197133376767018,
172
+ "eval_LAW_precision": 0.8558014944311293,
173
+ "eval_LAW_recall": 0.9939415771548337,
174
+ "eval_accuracy_normalized": 0.9867346072122146,
175
+ "eval_loss": 0.05432054027915001,
176
+ "eval_macro-f1": 0.944648370774704,
177
+ "eval_macro-precision": 0.9786239824274723,
178
+ "eval_macro-recall": 0.915818979128866,
179
+ "eval_micro-f1": 0.9336130116300778,
180
+ "eval_micro-precision": 0.9849840107521899,
181
+ "eval_micro-recall": 0.8873348224537085,
182
+ "eval_runtime": 182.9328,
183
+ "eval_samples_per_second": 69.616,
184
+ "eval_steps_per_second": 0.547,
185
+ "eval_weighted-f1": 0.93286298930164,
186
+ "eval_weighted-precision": 0.9859705314716058,
187
+ "eval_weighted-recall": 0.8873348224537085,
188
+ "step": 4224
189
+ },
190
+ {
191
+ "epoch": 7.0,
192
+ "learning_rate": 8.6e-06,
193
+ "loss": 0.0407,
194
+ "step": 4928
195
+ },
196
+ {
197
+ "epoch": 7.0,
198
+ "eval_CITATION_f1": 0.9689505360915142,
199
+ "eval_CITATION_precision": 0.9746612140165183,
200
+ "eval_CITATION_recall": 0.9633063877001109,
201
+ "eval_LAW_f1": 0.9199606090447693,
202
+ "eval_LAW_precision": 0.8559869181538808,
203
+ "eval_LAW_recall": 0.9942690594707886,
204
+ "eval_accuracy_normalized": 0.9867467947680346,
205
+ "eval_loss": 0.04946836084127426,
206
+ "eval_macro-f1": 0.94497928086637,
207
+ "eval_macro-precision": 0.9787877235854497,
208
+ "eval_macro-recall": 0.9163274179869876,
209
+ "eval_micro-f1": 0.9339147052685917,
210
+ "eval_micro-precision": 0.9852157389813228,
211
+ "eval_micro-recall": 0.8876918258690887,
212
+ "eval_runtime": 183.49,
213
+ "eval_samples_per_second": 69.404,
214
+ "eval_steps_per_second": 0.545,
215
+ "eval_weighted-f1": 0.9331590580868986,
216
+ "eval_weighted-precision": 0.9862179243996534,
217
+ "eval_weighted-recall": 0.8876918258690887,
218
+ "step": 4928
219
+ },
220
+ {
221
+ "epoch": 8.0,
222
+ "learning_rate": 8.400000000000001e-06,
223
+ "loss": 0.0371,
224
+ "step": 5632
225
+ },
226
+ {
227
+ "epoch": 8.0,
228
+ "eval_CITATION_f1": 0.9696075303126994,
229
+ "eval_CITATION_precision": 0.975991649269311,
230
+ "eval_CITATION_recall": 0.9633063877001109,
231
+ "eval_LAW_f1": 0.9197064830925183,
232
+ "eval_LAW_precision": 0.85474844624427,
233
+ "eval_LAW_recall": 0.9953497511134399,
234
+ "eval_accuracy_normalized": 0.9866988062669936,
235
+ "eval_loss": 0.045815542340278625,
236
+ "eval_macro-f1": 0.9452922817050826,
237
+ "eval_macro-precision": 0.9793280694067754,
238
+ "eval_macro-recall": 0.9165665634793729,
239
+ "eval_micro-f1": 0.934054837222576,
240
+ "eval_micro-precision": 0.9859804421374612,
241
+ "eval_micro-recall": 0.8873248248248248,
242
+ "eval_runtime": 182.9672,
243
+ "eval_samples_per_second": 69.603,
244
+ "eval_steps_per_second": 0.547,
245
+ "eval_weighted-f1": 0.9332733879110633,
246
+ "eval_weighted-precision": 0.9870388712591779,
247
+ "eval_weighted-recall": 0.8873248248248248,
248
+ "step": 5632
249
+ },
250
+ {
251
+ "epoch": 9.0,
252
+ "learning_rate": 8.2e-06,
253
+ "loss": 0.0345,
254
+ "step": 6336
255
+ },
256
+ {
257
+ "epoch": 9.0,
258
+ "eval_CITATION_f1": 0.9796242602375184,
259
+ "eval_CITATION_precision": 0.9819253125248826,
260
+ "eval_CITATION_recall": 0.9773339673482326,
261
+ "eval_LAW_f1": 0.9201594688414254,
262
+ "eval_LAW_precision": 0.8565743798154263,
263
+ "eval_LAW_recall": 0.9939415771548337,
264
+ "eval_accuracy_normalized": 0.9868583870760111,
265
+ "eval_loss": 0.04324862360954285,
266
+ "eval_macro-f1": 0.9502039808194804,
267
+ "eval_macro-precision": 0.9856377722515332,
268
+ "eval_macro-recall": 0.9198232236455686,
269
+ "eval_micro-f1": 0.9369347616120818,
270
+ "eval_micro-precision": 0.9890856004078417,
271
+ "eval_micro-recall": 0.8900079235998165,
272
+ "eval_runtime": 182.6619,
273
+ "eval_samples_per_second": 69.719,
274
+ "eval_steps_per_second": 0.547,
275
+ "eval_weighted-f1": 0.9360739408863803,
276
+ "eval_weighted-precision": 0.9895945375388725,
277
+ "eval_weighted-recall": 0.8900079235998165,
278
+ "step": 6336
279
+ },
280
+ {
281
+ "epoch": 10.0,
282
+ "learning_rate": 8.000000000000001e-06,
283
+ "loss": 0.0326,
284
+ "step": 7040
285
+ },
286
+ {
287
+ "epoch": 10.0,
288
+ "eval_CITATION_f1": 0.98077534159517,
289
+ "eval_CITATION_precision": 0.9831183309444179,
290
+ "eval_CITATION_recall": 0.9784434934220955,
291
+ "eval_LAW_f1": 0.9204929588139884,
292
+ "eval_LAW_precision": 0.8568848248807609,
293
+ "eval_LAW_recall": 0.9943018077023841,
294
+ "eval_accuracy_normalized": 0.9868957114657098,
295
+ "eval_loss": 0.04360693693161011,
296
+ "eval_macro-f1": 0.9508766633650136,
297
+ "eval_macro-precision": 0.9863726505622398,
298
+ "eval_macro-recall": 0.9204543086360064,
299
+ "eval_micro-f1": 0.9373909417148626,
300
+ "eval_micro-precision": 0.9896649209806738,
301
+ "eval_micro-recall": 0.8903621239602226,
302
+ "eval_runtime": 182.4045,
303
+ "eval_samples_per_second": 69.817,
304
+ "eval_steps_per_second": 0.548,
305
+ "eval_weighted-f1": 0.9365143149914062,
306
+ "eval_weighted-precision": 0.9901520055290943,
307
+ "eval_weighted-recall": 0.8903621239602226,
308
+ "step": 7040
309
+ },
310
+ {
311
+ "epoch": 11.0,
312
+ "learning_rate": 7.800000000000002e-06,
313
+ "loss": 0.0312,
314
+ "step": 7744
315
+ },
316
+ {
317
+ "epoch": 11.0,
318
+ "eval_CITATION_f1": 0.9814785373608903,
319
+ "eval_CITATION_precision": 0.9844522404720141,
320
+ "eval_CITATION_recall": 0.9785227452845142,
321
+ "eval_LAW_f1": 0.920169278140974,
322
+ "eval_LAW_precision": 0.8570540532903845,
323
+ "eval_LAW_recall": 0.9933193607545192,
324
+ "eval_accuracy_normalized": 0.9868713363540699,
325
+ "eval_loss": 0.041282109916210175,
326
+ "eval_macro-f1": 0.9510066434146158,
327
+ "eval_macro-precision": 0.9859210530195167,
328
+ "eval_macro-recall": 0.9210918842900306,
329
+ "eval_micro-f1": 0.937299600298678,
330
+ "eval_micro-precision": 0.9889929091161885,
331
+ "eval_micro-recall": 0.8907417456275828,
332
+ "eval_runtime": 183.3833,
333
+ "eval_samples_per_second": 69.445,
334
+ "eval_steps_per_second": 0.545,
335
+ "eval_weighted-f1": 0.936407960886005,
336
+ "eval_weighted-precision": 0.9894474242690886,
337
+ "eval_weighted-recall": 0.8907417456275828,
338
+ "step": 7744
339
+ },
340
+ {
341
+ "epoch": 12.0,
342
+ "learning_rate": 7.600000000000001e-06,
343
+ "loss": 0.0302,
344
+ "step": 8448
345
+ },
346
+ {
347
+ "epoch": 12.0,
348
+ "eval_CITATION_f1": 0.9821840451761711,
349
+ "eval_CITATION_precision": 0.9857120051085568,
350
+ "eval_CITATION_recall": 0.9786812490093517,
351
+ "eval_LAW_f1": 0.9202834903231668,
352
+ "eval_LAW_precision": 0.8559637162657051,
353
+ "eval_LAW_recall": 0.9950550170290804,
354
+ "eval_accuracy_normalized": 0.9868378205755649,
355
+ "eval_loss": 0.0409509539604187,
356
+ "eval_macro-f1": 0.9515210413529415,
357
+ "eval_macro-precision": 0.986868133019216,
358
+ "eval_macro-recall": 0.9213572734980617,
359
+ "eval_micro-f1": 0.937704342483488,
360
+ "eval_micro-precision": 0.9902674143764193,
361
+ "eval_micro-recall": 0.8904400733455576,
362
+ "eval_runtime": 182.6717,
363
+ "eval_samples_per_second": 69.715,
364
+ "eval_steps_per_second": 0.547,
365
+ "eval_weighted-f1": 0.9367869766694694,
366
+ "eval_weighted-precision": 0.9907821159064123,
367
+ "eval_weighted-recall": 0.8904400733455576,
368
+ "step": 8448
369
+ },
370
+ {
371
+ "epoch": 13.0,
372
+ "learning_rate": 7.4e-06,
373
+ "loss": 0.0294,
374
+ "step": 9152
375
+ },
376
+ {
377
+ "epoch": 13.0,
378
+ "eval_CITATION_f1": 0.9822635806887775,
379
+ "eval_CITATION_precision": 0.9857918263090677,
380
+ "eval_CITATION_recall": 0.9787605008717705,
381
+ "eval_LAW_f1": 0.9207781394292636,
382
+ "eval_LAW_precision": 0.8561896449787437,
383
+ "eval_LAW_recall": 0.9959064710505633,
384
+ "eval_accuracy_normalized": 0.986947889439064,
385
+ "eval_loss": 0.039835721254348755,
386
+ "eval_macro-f1": 0.9517621380923427,
387
+ "eval_macro-precision": 0.9873334859611669,
388
+ "eval_macro-recall": 0.9214192450961509,
389
+ "eval_micro-f1": 0.9380497970823737,
390
+ "eval_micro-precision": 0.9908930805950781,
391
+ "eval_micro-recall": 0.8905573142285905,
392
+ "eval_runtime": 183.2667,
393
+ "eval_samples_per_second": 69.489,
394
+ "eval_steps_per_second": 0.546,
395
+ "eval_weighted-f1": 0.937137235367138,
396
+ "eval_weighted-precision": 0.9914335782136015,
397
+ "eval_weighted-recall": 0.8905573142285905,
398
+ "step": 9152
399
+ },
400
+ {
401
+ "epoch": 14.0,
402
+ "learning_rate": 7.2000000000000005e-06,
403
+ "loss": 0.0288,
404
+ "step": 9856
405
+ },
406
+ {
407
+ "epoch": 14.0,
408
+ "eval_CITATION_f1": 0.9823417117403754,
409
+ "eval_CITATION_precision": 0.9859492256107297,
410
+ "eval_CITATION_recall": 0.9787605008717705,
411
+ "eval_LAW_f1": 0.9209932279909707,
412
+ "eval_LAW_precision": 0.8569253756589891,
413
+ "eval_LAW_recall": 0.9954152475766309,
414
+ "eval_accuracy_normalized": 0.986973407134062,
415
+ "eval_loss": 0.03937051072716713,
416
+ "eval_macro-f1": 0.9518572703321915,
417
+ "eval_macro-precision": 0.9870878742242006,
418
+ "eval_macro-recall": 0.9217881303971995,
419
+ "eval_micro-f1": 0.9381220440903755,
420
+ "eval_micro-precision": 0.9905454882513788,
421
+ "eval_micro-recall": 0.8909685891156179,
422
+ "eval_runtime": 183.2841,
423
+ "eval_samples_per_second": 69.482,
424
+ "eval_steps_per_second": 0.546,
425
+ "eval_weighted-f1": 0.937211000059121,
426
+ "eval_weighted-precision": 0.9910683576452416,
427
+ "eval_weighted-recall": 0.8909685891156179,
428
+ "step": 9856
429
+ },
430
+ {
431
+ "epoch": 15.0,
432
+ "learning_rate": 7.000284090909092e-06,
433
+ "loss": 0.0282,
434
+ "step": 10560
435
+ },
436
+ {
437
+ "epoch": 15.0,
438
+ "eval_CITATION_f1": 0.9819121447028423,
439
+ "eval_CITATION_precision": 0.9850841509132966,
440
+ "eval_CITATION_recall": 0.9787605008717705,
441
+ "eval_LAW_f1": 0.9207796338083628,
442
+ "eval_LAW_precision": 0.8564586561487534,
443
+ "eval_LAW_recall": 0.9955462405030129,
444
+ "eval_accuracy_normalized": 0.9870015908568956,
445
+ "eval_loss": 0.039993565529584885,
446
+ "eval_macro-f1": 0.9517671412400261,
447
+ "eval_macro-precision": 0.9871533706873916,
448
+ "eval_macro-recall": 0.9215724095097284,
449
+ "eval_micro-f1": 0.9380245532040943,
450
+ "eval_micro-precision": 0.9906381795430319,
451
+ "eval_micro-recall": 0.8907177831024065,
452
+ "eval_runtime": 183.1632,
453
+ "eval_samples_per_second": 69.528,
454
+ "eval_steps_per_second": 0.546,
455
+ "eval_weighted-f1": 0.937111701553557,
456
+ "eval_weighted-precision": 0.9911661050067804,
457
+ "eval_weighted-recall": 0.8907177831024065,
458
+ "step": 10560
459
+ },
460
+ {
461
+ "epoch": 16.0,
462
+ "learning_rate": 6.800284090909092e-06,
463
+ "loss": 0.0278,
464
+ "step": 11264
465
+ },
466
+ {
467
+ "epoch": 16.0,
468
+ "eval_CITATION_f1": 0.9828889773179467,
469
+ "eval_CITATION_precision": 0.9870524296675192,
470
+ "eval_CITATION_recall": 0.9787605008717705,
471
+ "eval_LAW_f1": 0.9207239434911781,
472
+ "eval_LAW_precision": 0.8570662603002597,
473
+ "eval_LAW_recall": 0.9945965417867435,
474
+ "eval_accuracy_normalized": 0.9870153018571931,
475
+ "eval_loss": 0.03851037099957466,
476
+ "eval_macro-f1": 0.9520466388916144,
477
+ "eval_macro-precision": 0.986678521329257,
478
+ "eval_macro-recall": 0.9224923590174634,
479
+ "eval_micro-f1": 0.9381704786270355,
480
+ "eval_micro-precision": 0.9899661676785466,
481
+ "eval_micro-recall": 0.8915252822471254,
482
+ "eval_runtime": 183.6471,
483
+ "eval_samples_per_second": 69.345,
484
+ "eval_steps_per_second": 0.545,
485
+ "eval_weighted-f1": 0.937250674098129,
486
+ "eval_weighted-precision": 0.990462627560541,
487
+ "eval_weighted-recall": 0.8915252822471254,
488
+ "step": 11264
489
+ },
490
+ {
491
+ "epoch": 17.0,
492
+ "learning_rate": 6.6002840909090915e-06,
493
+ "loss": 0.0275,
494
+ "step": 11968
495
+ },
496
+ {
497
+ "epoch": 17.0,
498
+ "eval_CITATION_f1": 0.9827716547964827,
499
+ "eval_CITATION_precision": 0.9868158210147823,
500
+ "eval_CITATION_recall": 0.9787605008717705,
501
+ "eval_LAW_f1": 0.9203515684194575,
502
+ "eval_LAW_precision": 0.8565183054098268,
503
+ "eval_LAW_recall": 0.9944655488603615,
504
+ "eval_accuracy_normalized": 0.9869817860786881,
505
+ "eval_loss": 0.03849661350250244,
506
+ "eval_macro-f1": 0.9518505107642844,
507
+ "eval_macro-precision": 0.9866130248660661,
508
+ "eval_macro-recall": 0.9221843554847858,
509
+ "eval_micro-f1": 0.9379384544446518,
510
+ "eval_micro-precision": 0.9898734763868935,
511
+ "eval_micro-recall": 0.8911814407610623,
512
+ "eval_runtime": 182.5915,
513
+ "eval_samples_per_second": 69.746,
514
+ "eval_steps_per_second": 0.548,
515
+ "eval_weighted-f1": 0.9370160495313582,
516
+ "eval_weighted-precision": 0.990366044125977,
517
+ "eval_weighted-recall": 0.8911814407610623,
518
+ "step": 11968
519
+ },
520
+ {
521
+ "epoch": 18.0,
522
+ "learning_rate": 6.400568181818182e-06,
523
+ "loss": 0.0272,
524
+ "step": 12672
525
+ },
526
+ {
527
+ "epoch": 18.0,
528
+ "eval_CITATION_f1": 0.9828889773179467,
529
+ "eval_CITATION_precision": 0.9870524296675192,
530
+ "eval_CITATION_recall": 0.9787605008717705,
531
+ "eval_LAW_f1": 0.9205657006147603,
532
+ "eval_LAW_precision": 0.8561369909311103,
533
+ "eval_LAW_recall": 0.9954807440398219,
534
+ "eval_accuracy_normalized": 0.9869932119122694,
535
+ "eval_loss": 0.038164205849170685,
536
+ "eval_macro-f1": 0.9519616843549739,
537
+ "eval_macro-precision": 0.9871206224557962,
538
+ "eval_macro-recall": 0.9220114522283747,
539
+ "eval_micro-f1": 0.9380527089596453,
540
+ "eval_micro-precision": 0.9905918338972054,
541
+ "eval_micro-recall": 0.8908060348420438,
542
+ "eval_runtime": 182.7469,
543
+ "eval_samples_per_second": 69.687,
544
+ "eval_steps_per_second": 0.547,
545
+ "eval_weighted-f1": 0.9371218313775519,
546
+ "eval_weighted-precision": 0.9911219409633794,
547
+ "eval_weighted-recall": 0.8908060348420438,
548
+ "step": 12672
549
+ },
550
+ {
551
+ "epoch": 19.0,
552
+ "learning_rate": 6.200568181818183e-06,
553
+ "loss": 0.027,
554
+ "step": 13376
555
+ },
556
+ {
557
+ "epoch": 19.0,
558
+ "eval_CITATION_f1": 0.982810759191469,
559
+ "eval_CITATION_precision": 0.9868946779606841,
560
+ "eval_CITATION_recall": 0.9787605008717705,
561
+ "eval_LAW_f1": 0.9201540514313441,
562
+ "eval_LAW_precision": 0.8567596566523605,
563
+ "eval_LAW_recall": 0.9936795913020697,
564
+ "eval_accuracy_normalized": 0.986947889439064,
565
+ "eval_loss": 0.03846826031804085,
566
+ "eval_macro-f1": 0.9517588185840318,
567
+ "eval_macro-precision": 0.9862200460869202,
568
+ "eval_macro-recall": 0.922317690725456,
569
+ "eval_micro-f1": 0.9378130217028381,
570
+ "eval_micro-precision": 0.9893173286369745,
571
+ "eval_micro-recall": 0.891406021631102,
572
+ "eval_runtime": 184.1158,
573
+ "eval_samples_per_second": 69.168,
574
+ "eval_steps_per_second": 0.543,
575
+ "eval_weighted-f1": 0.9368937680506739,
576
+ "eval_weighted-precision": 0.9897820747141067,
577
+ "eval_weighted-recall": 0.891406021631102,
578
+ "step": 13376
579
+ },
580
+ {
581
+ "epoch": 20.0,
582
+ "learning_rate": 6.0005681818181824e-06,
583
+ "loss": 0.0266,
584
+ "step": 14080
585
+ },
586
+ {
587
+ "epoch": 20.0,
588
+ "eval_CITATION_f1": 0.982928091050181,
589
+ "eval_CITATION_precision": 0.9871313244344977,
590
+ "eval_CITATION_recall": 0.9787605008717705,
591
+ "eval_LAW_f1": 0.9199981821211617,
592
+ "eval_LAW_precision": 0.8559548978153629,
593
+ "eval_LAW_recall": 0.9944000523971706,
594
+ "eval_accuracy_normalized": 0.9869730262729426,
595
+ "eval_loss": 0.03889286145567894,
596
+ "eval_macro-f1": 0.9518160523849399,
597
+ "eval_macro-precision": 0.9865802766344705,
598
+ "eval_macro-recall": 0.9221656032832765,
599
+ "eval_micro-f1": 0.9378327643177853,
600
+ "eval_micro-precision": 0.9898271307410669,
601
+ "eval_micro-recall": 0.8910281816475104,
602
+ "eval_runtime": 183.0494,
603
+ "eval_samples_per_second": 69.571,
604
+ "eval_steps_per_second": 0.546,
605
+ "eval_weighted-f1": 0.936903094166417,
606
+ "eval_weighted-precision": 0.9903191339376443,
607
+ "eval_weighted-recall": 0.8910281816475104,
608
+ "step": 14080
609
+ },
610
+ {
611
+ "epoch": 21.0,
612
+ "learning_rate": 5.800568181818183e-06,
613
+ "loss": 0.0264,
614
+ "step": 14784
615
+ },
616
+ {
617
+ "epoch": 21.0,
618
+ "eval_CITATION_f1": 0.9821030862233535,
619
+ "eval_CITATION_precision": 0.9857097237745489,
620
+ "eval_CITATION_recall": 0.9785227452845142,
621
+ "eval_LAW_f1": 0.9203370820576556,
622
+ "eval_LAW_precision": 0.856639015856893,
623
+ "eval_LAW_recall": 0.9942690594707886,
624
+ "eval_accuracy_normalized": 0.9869707411062264,
625
+ "eval_loss": 0.038498904556035995,
626
+ "eval_macro-f1": 0.9515075473026433,
627
+ "eval_macro-precision": 0.9863959023776514,
628
+ "eval_macro-recall": 0.9216946659139149,
629
+ "eval_micro-f1": 0.9377305462849113,
630
+ "eval_micro-precision": 0.9896649209806738,
631
+ "eval_micro-recall": 0.890975090749781,
632
+ "eval_runtime": 182.7777,
633
+ "eval_samples_per_second": 69.675,
634
+ "eval_steps_per_second": 0.547,
635
+ "eval_weighted-f1": 0.9368225906951636,
636
+ "eval_weighted-precision": 0.9901555837282984,
637
+ "eval_weighted-recall": 0.890975090749781,
638
+ "step": 14784
639
+ },
640
+ {
641
+ "epoch": 22.0,
642
+ "learning_rate": 5.600568181818182e-06,
643
+ "loss": 0.0262,
644
+ "step": 15488
645
+ },
646
+ {
647
+ "epoch": 22.0,
648
+ "eval_CITATION_f1": 0.9823403070559223,
649
+ "eval_CITATION_precision": 0.9860268284893006,
650
+ "eval_CITATION_recall": 0.9786812490093517,
651
+ "eval_LAW_f1": 0.9201017657570342,
652
+ "eval_LAW_precision": 0.8557946926587413,
653
+ "eval_LAW_recall": 0.9948585276395074,
654
+ "eval_accuracy_normalized": 0.986917420549514,
655
+ "eval_loss": 0.040023379027843475,
656
+ "eval_macro-f1": 0.951632458392357,
657
+ "eval_macro-precision": 0.9867698883244296,
658
+ "eval_macro-recall": 0.9216507907951708,
659
+ "eval_micro-f1": 0.9377784605930252,
660
+ "eval_micro-precision": 0.9901283774389397,
661
+ "eval_micro-recall": 0.8906862336362878,
662
+ "eval_runtime": 182.7513,
663
+ "eval_samples_per_second": 69.685,
664
+ "eval_steps_per_second": 0.547,
665
+ "eval_weighted-f1": 0.9368585268788033,
666
+ "eval_weighted-precision": 0.9906368284989479,
667
+ "eval_weighted-recall": 0.8906862336362878,
668
+ "step": 15488
669
+ },
670
+ {
671
+ "epoch": 23.0,
672
+ "learning_rate": 5.400568181818182e-06,
673
+ "loss": 0.0259,
674
+ "step": 16192
675
+ },
676
+ {
677
+ "epoch": 23.0,
678
+ "eval_CITATION_f1": 0.9820626019170345,
679
+ "eval_CITATION_precision": 0.9857085828343314,
680
+ "eval_CITATION_recall": 0.9784434934220955,
681
+ "eval_LAW_f1": 0.9201865140642503,
682
+ "eval_LAW_precision": 0.8556506560054057,
683
+ "eval_LAW_recall": 0.9952515064186533,
684
+ "eval_accuracy_normalized": 0.9869360827443634,
685
+ "eval_loss": 0.03981756791472435,
686
+ "eval_macro-f1": 0.9515610106395767,
687
+ "eval_macro-precision": 0.9868474999203745,
688
+ "eval_macro-recall": 0.9214737971120037,
689
+ "eval_micro-f1": 0.9377598823877911,
690
+ "eval_micro-precision": 0.9903369328451592,
691
+ "eval_micro-recall": 0.8904840289208843,
692
+ "eval_runtime": 182.5734,
693
+ "eval_samples_per_second": 69.753,
694
+ "eval_steps_per_second": 0.548,
695
+ "eval_weighted-f1": 0.9368400401851952,
696
+ "eval_weighted-precision": 0.9908674773583548,
697
+ "eval_weighted-recall": 0.8904840289208843,
698
+ "step": 16192
699
+ },
700
+ {
701
+ "epoch": 23.0,
702
+ "step": 16192,
703
+ "total_flos": 1.3557575111977882e+17,
704
+ "train_loss": 0.05992582664188189,
705
+ "train_runtime": 10758.5934,
706
+ "train_samples_per_second": 418.74,
707
+ "train_steps_per_second": 3.272
708
+ },
709
+ {
710
+ "epoch": 23.0,
711
+ "eval_CITATION_f1": 0.9814787122684838,
712
+ "eval_CITATION_precision": 0.9844650624611626,
713
+ "eval_CITATION_recall": 0.978510425318936,
714
+ "eval_LAW_f1": 0.9211885809936669,
715
+ "eval_LAW_precision": 0.8550646934923544,
716
+ "eval_LAW_recall": 0.9983966169255928,
717
+ "eval_accuracy_normalized": 0.9905667906095162,
718
+ "eval_loss": 0.026775669306516647,
719
+ "eval_macro-f1": 0.9515974221562933,
720
+ "eval_macro-precision": 0.9884535211222644,
721
+ "eval_macro-recall": 0.9202279237061262,
722
+ "eval_micro-f1": 0.9429566347321088,
723
+ "eval_micro-precision": 0.9909403074465961,
724
+ "eval_micro-recall": 0.8994053033460294,
725
+ "eval_runtime": 1297.076,
726
+ "eval_samples_per_second": 69.465,
727
+ "eval_steps_per_second": 0.543,
728
+ "eval_weighted-f1": 0.9419101385291693,
729
+ "eval_weighted-precision": 0.9916707876514618,
730
+ "eval_weighted-recall": 0.8994053033460294,
731
+ "step": 16192
732
+ }
733
+ ],
734
+ "max_steps": 35200,
735
+ "num_train_epochs": 50,
736
+ "total_flos": 1.3557575111977882e+17,
737
+ "trial_name": null,
738
+ "trial_params": null
739
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f786e36e587f57aa16e36e866bdbaa5961b59563fddb5e05c500499e1282f912
3
+ size 3707