asahi417 commited on
Commit
932c32b
1 Parent(s): abca2f4

model update

Browse files
README.md ADDED
@@ -0,0 +1,176 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ datasets:
3
+ - tner/tweetner7
4
+ metrics:
5
+ - f1
6
+ - precision
7
+ - recall
8
+ model-index:
9
+ - name: tner/bertweet-base-tweetner7-2021
10
+ results:
11
+ - task:
12
+ name: Token Classification
13
+ type: token-classification
14
+ dataset:
15
+ name: tner/tweetner7/test_2021
16
+ type: tner/tweetner7/test_2021
17
+ args: tner/tweetner7/test_2021
18
+ metrics:
19
+ - name: F1
20
+ type: f1
21
+ value: 0.6308962917798349
22
+ - name: Precision
23
+ type: precision
24
+ value: 0.6058767167039285
25
+ - name: Recall
26
+ type: recall
27
+ value: 0.6580712303422757
28
+ - name: F1 (macro)
29
+ type: f1_macro
30
+ value: 0.5735468406550763
31
+ - name: Precision (macro)
32
+ type: precision_macro
33
+ value: 0.5503198173085064
34
+ - name: Recall (macro)
35
+ type: recall_macro
36
+ value: 0.6012922054817469
37
+ - name: F1 (entity span)
38
+ type: f1_entity_span
39
+ value: 0.7788214245778822
40
+ - name: Precision (entity span)
41
+ type: precision_entity_span
42
+ value: 0.7538694663924668
43
+ - name: Recall (entity span)
44
+ type: recall_entity_span
45
+ value: 0.8054816699433329
46
+ - task:
47
+ name: Token Classification
48
+ type: token-classification
49
+ dataset:
50
+ name: tner/tweetner7/test_2020
51
+ type: tner/tweetner7/test_2020
52
+ args: tner/tweetner7/test_2020
53
+ metrics:
54
+ - name: F1
55
+ type: f1
56
+ value: 0.6205787781350482
57
+ - name: Precision
58
+ type: precision
59
+ value: 0.6415512465373961
60
+ - name: Recall
61
+ type: recall
62
+ value: 0.6009340944473275
63
+ - name: F1 (macro)
64
+ type: f1_macro
65
+ value: 0.5723158793505982
66
+ - name: Precision (macro)
67
+ type: precision_macro
68
+ value: 0.5910271170769507
69
+ - name: Recall (macro)
70
+ type: recall_macro
71
+ value: 0.5568451570610017
72
+ - name: F1 (entity span)
73
+ type: f1_entity_span
74
+ value: 0.7595141700404859
75
+ - name: Precision (entity span)
76
+ type: precision_entity_span
77
+ value: 0.7913385826771654
78
+ - name: Recall (entity span)
79
+ type: recall_entity_span
80
+ value: 0.7301504929942917
81
+
82
+ pipeline_tag: token-classification
83
+ widget:
84
+ - text: "Get the all-analog Classic Vinyl Edition of `Takin' Off` Album from {{@Herbie Hancock@}} via {{USERNAME}} link below: {{URL}}"
85
+ example_title: "NER Example 1"
86
+ ---
87
+ # tner/bertweet-base-tweetner7-2021
88
+
89
+ This model is a fine-tuned version of [vinai/bertweet-base](https://huggingface.co/vinai/bertweet-base) on the
90
+ [tner/tweetner7](https://huggingface.co/datasets/tner/tweetner7) dataset (`train_2021` split).
91
+ Model fine-tuning is done via [T-NER](https://github.com/asahi417/tner)'s hyper-parameter search (see the repository
92
+ for more detail). It achieves the following results on the test set of 2021:
93
+ - F1 (micro): 0.6308962917798349
94
+ - Precision (micro): 0.6058767167039285
95
+ - Recall (micro): 0.6580712303422757
96
+ - F1 (macro): 0.5735468406550763
97
+ - Precision (macro): 0.5503198173085064
98
+ - Recall (macro): 0.6012922054817469
99
+
100
+
101
+
102
+ The per-entity breakdown of the F1 score on the test set are below:
103
+ - corporation: 0.4565701559020044
104
+ - creative_work: 0.4098984771573604
105
+ - event: 0.4628410159924742
106
+ - group: 0.593177511054959
107
+ - location: 0.6333949476278496
108
+ - person: 0.8279457768508863
109
+ - product: 0.631
110
+
111
+ For F1 scores, the confidence interval is obtained by bootstrap as below:
112
+ - F1 (micro):
113
+ - 90%: [0.6218627510838193, 0.6407164862470697]
114
+ - 95%: [0.6201627010426306, 0.6422908401462293]
115
+ - F1 (macro):
116
+ - 90%: [0.6218627510838193, 0.6407164862470697]
117
+ - 95%: [0.6201627010426306, 0.6422908401462293]
118
+
119
+ Full evaluation can be found at [metric file of NER](https://huggingface.co/tner/bertweet-base-tweetner7-2021/raw/main/eval/metric.json)
120
+ and [metric file of entity span](https://huggingface.co/tner/bertweet-base-tweetner7-2021/raw/main/eval/metric_span.json).
121
+
122
+ ### Usage
123
+ This model can be used through the [tner library](https://github.com/asahi417/tner). Install the library via pip
124
+ ```shell
125
+ pip install tner
126
+ ```
127
+ and activate model as below.
128
+ ```python
129
+ from tner import TransformersNER
130
+ model = TransformersNER("tner/bertweet-base-tweetner7-2021")
131
+ model.predict(["Jacob Collier is a Grammy awarded English artist from London"])
132
+ ```
133
+ It can be used via transformers library but it is not recommended as CRF layer is not supported at the moment.
134
+
135
+ ### Training hyperparameters
136
+
137
+ The following hyperparameters were used during training:
138
+ - dataset: ['tner/tweetner7']
139
+ - dataset_split: train_2021
140
+ - dataset_name: None
141
+ - local_dataset: None
142
+ - model: vinai/bertweet-base
143
+ - crf: False
144
+ - max_length: 128
145
+ - epoch: 30
146
+ - batch_size: 32
147
+ - lr: 0.0001
148
+ - random_seed: 0
149
+ - gradient_accumulation_steps: 1
150
+ - weight_decay: 1e-07
151
+ - lr_warmup_step_ratio: 0.3
152
+ - max_grad_norm: 1
153
+
154
+ The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/tner/bertweet-base-tweetner7-2021/raw/main/trainer_config.json).
155
+
156
+ ### Reference
157
+ If you use any resource from T-NER, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
158
+
159
+ ```
160
+
161
+ @inproceedings{ushio-camacho-collados-2021-ner,
162
+ title = "{T}-{NER}: An All-Round Python Library for Transformer-based Named Entity Recognition",
163
+ author = "Ushio, Asahi and
164
+ Camacho-Collados, Jose",
165
+ booktitle = "Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations",
166
+ month = apr,
167
+ year = "2021",
168
+ address = "Online",
169
+ publisher = "Association for Computational Linguistics",
170
+ url = "https://aclanthology.org/2021.eacl-demos.7",
171
+ doi = "10.18653/v1/2021.eacl-demos.7",
172
+ pages = "53--62",
173
+ abstract = "Language model (LM) pretraining has led to consistent improvements in many NLP downstream tasks, including named entity recognition (NER). In this paper, we present T-NER (Transformer-based Named Entity Recognition), a Python library for NER LM finetuning. In addition to its practical utility, T-NER facilitates the study and investigation of the cross-domain and cross-lingual generalization ability of LMs finetuned on NER. Our library also provides a web app where users can get model predictions interactively for arbitrary text, which facilitates qualitative model evaluation for non-expert programmers. We show the potential of the library by compiling nine public NER datasets into a unified format and evaluating the cross-domain and cross- lingual performance across the datasets. The results from our initial experiments show that in-domain performance is generally competitive across datasets. However, cross-domain generalization is challenging even with a large pretrained LM, which has nevertheless capacity to learn domain-specific features if fine- tuned on a combined dataset. To facilitate future research, we also release all our LM checkpoints via the Hugging Face model hub.",
174
+ }
175
+
176
+ ```
eval/metric.json DELETED
@@ -1 +0,0 @@
1
- {"2021.dev": {"micro/f1": 0.6220703125000001, "micro/f1_ci": {}, "micro/recall": 0.637, "micro/precision": 0.607824427480916, "macro/f1": 0.5734483206270202, "macro/f1_ci": {}, "macro/recall": 0.5876986774163436, "macro/precision": 0.5610235583875315, "per_entity_metric": {"corporation": {"f1": 0.5217391304347826, "f1_ci": {}, "precision": 0.5142857142857142, "recall": 0.5294117647058824}, "creative_work": {"f1": 0.45333333333333337, "f1_ci": {}, "precision": 0.4473684210526316, "recall": 0.4594594594594595}, "event": {"f1": 0.38247011952191234, "f1_ci": {}, "precision": 0.4, "recall": 0.366412213740458}, "group": {"f1": 0.6214442013129103, "f1_ci": {}, "precision": 0.6173913043478261, "recall": 0.6255506607929515}, "location": {"f1": 0.6184210526315789, "f1_ci": {}, "precision": 0.5875, "recall": 0.6527777777777778}, "person": {"f1": 0.8006589785831961, "f1_ci": {}, "precision": 0.75, "recall": 0.8586572438162544}, "product": {"f1": 0.6160714285714285, "f1_ci": {}, "precision": 0.6106194690265486, "recall": 0.6216216216216216}}}, "2021.test": {"micro/f1": 0.6308962917798349, "micro/f1_ci": {"90": [0.6218627510838193, 0.6407164862470697], "95": [0.6201627010426306, 0.6422908401462293]}, "micro/recall": 0.6580712303422757, "micro/precision": 0.6058767167039285, "macro/f1": 0.5735468406550763, "macro/f1_ci": {"90": [0.5633518872919319, 0.5834141537051213], "95": [0.5619155878559741, 0.585359923196132]}, "macro/recall": 0.6012922054817469, "macro/precision": 0.5503198173085064, "per_entity_metric": {"corporation": {"f1": 0.4565701559020044, "f1_ci": {"90": [0.43122213480199306, 0.4842832655545205], "95": [0.422987517373063, 0.4892101769228394]}, "precision": 0.4575892857142857, "recall": 0.45555555555555555}, "creative_work": {"f1": 0.4098984771573604, "f1_ci": {"90": [0.3791530120969373, 0.4415152462146761], "95": [0.3716394651460033, 0.4462899138884782]}, "precision": 0.38224852071005916, "recall": 0.4418604651162791}, "event": {"f1": 0.4628410159924742, "f1_ci": {"90": [0.43764436774212423, 0.4844905070216916], "95": [0.4327500906125407, 0.49103704697586353]}, "precision": 0.4790652385589094, "recall": 0.44767970882620567}, "group": {"f1": 0.593177511054959, "f1_ci": {"90": [0.5717044673748471, 0.6144093322442311], "95": [0.5676831336909995, 0.6175847266352983]}, "precision": 0.5697815533980582, "recall": 0.6185770750988142}, "location": {"f1": 0.6333949476278496, "f1_ci": {"90": [0.6061374762309342, 0.6596729625322199], "95": [0.5999818181818181, 0.6642491807610421]}, "precision": 0.5667034178610805, "recall": 0.7178770949720671}, "person": {"f1": 0.8279457768508863, "f1_ci": {"90": [0.8177849742494966, 0.8392861816686192], "95": [0.8161339468587502, 0.8409499904355973]}, "precision": 0.7830374753451677, "recall": 0.8783185840707964}, "product": {"f1": 0.631, "f1_ci": {"90": [0.6078329186866498, 0.6538099694036692], "95": [0.602711796196466, 0.6565176505572475]}, "precision": 0.6138132295719845, "recall": 0.6491769547325102}}}, "2020.test": {"micro/f1": 0.6205787781350482, "micro/f1_ci": {"90": [0.5990600374135929, 0.641028413028413], "95": [0.5956429008859256, 0.645577807233236]}, "micro/recall": 0.6009340944473275, "micro/precision": 0.6415512465373961, "macro/f1": 0.5723158793505982, "macro/f1_ci": {"90": [0.5491541277371618, 0.593824224484208], "95": [0.5446730617743423, 0.5989182517444881]}, "macro/recall": 0.5568451570610017, "macro/precision": 0.5910271170769507, "per_entity_metric": {"corporation": {"f1": 0.49867374005305043, "f1_ci": {"90": [0.4362597402597403, 0.5552962052962053], "95": [0.4225177366931115, 0.5675940646528883]}, "precision": 0.5053763440860215, "recall": 0.49214659685863876}, "creative_work": {"f1": 0.4583333333333333, "f1_ci": {"90": [0.3903903903903904, 0.516321444502628], "95": [0.3776601932639084, 0.5297937192118228]}, "precision": 0.49044585987261147, "recall": 0.4301675977653631}, "event": {"f1": 0.43892339544513453, "f1_ci": {"90": [0.38343874372946507, 0.4931036128685116], "95": [0.3732780599111177, 0.5052203054609383]}, "precision": 0.48623853211009177, "recall": 0.4}, "group": {"f1": 0.5264957264957265, "f1_ci": {"90": [0.4739659822849217, 0.5779008444686745], "95": [0.4623267883150051, 0.5874770558415155]}, "precision": 0.5620437956204379, "recall": 0.49517684887459806}, "location": {"f1": 0.6358381502890174, "f1_ci": {"90": [0.5734652877656781, 0.6918347805270832], "95": [0.5622911498701775, 0.7034883720930232]}, "precision": 0.6077348066298343, "recall": 0.6666666666666666}, "person": {"f1": 0.81787521079258, "f1_ci": {"90": [0.7927999152425893, 0.8405389863292992], "95": [0.7873498023715415, 0.844675509305848]}, "precision": 0.8220338983050848, "recall": 0.8137583892617449}, "product": {"f1": 0.630071599045346, "f1_ci": {"90": [0.5727563482336753, 0.683606306263845], "95": [0.5630568382452805, 0.6948397887323944]}, "precision": 0.6633165829145728, "recall": 0.6}}}, "2021.test (span detection)": {"micro/f1": 0.7788214245778822, "micro/f1_ci": {}, "micro/recall": 0.8054816699433329, "micro/precision": 0.7538694663924668, "macro/f1": 0.7788214245778822, "macro/f1_ci": {}, "macro/recall": 0.8054816699433329, "macro/precision": 0.7538694663924668}, "2020.test (span detection)": {"micro/f1": 0.7595141700404859, "micro/f1_ci": {}, "micro/recall": 0.7301504929942917, "micro/precision": 0.7913385826771654, "macro/f1": 0.7595141700404859, "macro/f1_ci": {}, "macro/recall": 0.7301504929942917, "macro/precision": 0.7913385826771654}}
 
 
eval/metric.test_2020.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"micro/f1": 0.6205787781350482, "micro/f1_ci": {"90": [0.5990600374135929, 0.641028413028413], "95": [0.5956429008859256, 0.645577807233236]}, "micro/recall": 0.6009340944473275, "micro/precision": 0.6415512465373961, "macro/f1": 0.5723158793505982, "macro/f1_ci": {"90": [0.5491541277371618, 0.593824224484208], "95": [0.5446730617743423, 0.5989182517444881]}, "macro/recall": 0.5568451570610017, "macro/precision": 0.5910271170769507, "per_entity_metric": {"corporation": {"f1": 0.49867374005305043, "f1_ci": {"90": [0.4362597402597403, 0.5552962052962053], "95": [0.4225177366931115, 0.5675940646528883]}, "precision": 0.5053763440860215, "recall": 0.49214659685863876}, "creative_work": {"f1": 0.4583333333333333, "f1_ci": {"90": [0.3903903903903904, 0.516321444502628], "95": [0.3776601932639084, 0.5297937192118228]}, "precision": 0.49044585987261147, "recall": 0.4301675977653631}, "event": {"f1": 0.43892339544513453, "f1_ci": {"90": [0.38343874372946507, 0.4931036128685116], "95": [0.3732780599111177, 0.5052203054609383]}, "precision": 0.48623853211009177, "recall": 0.4}, "group": {"f1": 0.5264957264957265, "f1_ci": {"90": [0.4739659822849217, 0.5779008444686745], "95": [0.4623267883150051, 0.5874770558415155]}, "precision": 0.5620437956204379, "recall": 0.49517684887459806}, "location": {"f1": 0.6358381502890174, "f1_ci": {"90": [0.5734652877656781, 0.6918347805270832], "95": [0.5622911498701775, 0.7034883720930232]}, "precision": 0.6077348066298343, "recall": 0.6666666666666666}, "person": {"f1": 0.81787521079258, "f1_ci": {"90": [0.7927999152425893, 0.8405389863292992], "95": [0.7873498023715415, 0.844675509305848]}, "precision": 0.8220338983050848, "recall": 0.8137583892617449}, "product": {"f1": 0.630071599045346, "f1_ci": {"90": [0.5727563482336753, 0.683606306263845], "95": [0.5630568382452805, 0.6948397887323944]}, "precision": 0.6633165829145728, "recall": 0.6}}}
eval/metric.test_2021.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"micro/f1": 0.6308962917798349, "micro/f1_ci": {"90": [0.6218627510838193, 0.6407164862470697], "95": [0.6201627010426306, 0.6422908401462293]}, "micro/recall": 0.6580712303422757, "micro/precision": 0.6058767167039285, "macro/f1": 0.5735468406550763, "macro/f1_ci": {"90": [0.5633518872919319, 0.5834141537051213], "95": [0.5619155878559741, 0.585359923196132]}, "macro/recall": 0.6012922054817469, "macro/precision": 0.5503198173085064, "per_entity_metric": {"corporation": {"f1": 0.4565701559020044, "f1_ci": {"90": [0.43122213480199306, 0.4842832655545205], "95": [0.422987517373063, 0.4892101769228394]}, "precision": 0.4575892857142857, "recall": 0.45555555555555555}, "creative_work": {"f1": 0.4098984771573604, "f1_ci": {"90": [0.3791530120969373, 0.4415152462146761], "95": [0.3716394651460033, 0.4462899138884782]}, "precision": 0.38224852071005916, "recall": 0.4418604651162791}, "event": {"f1": 0.4628410159924742, "f1_ci": {"90": [0.43764436774212423, 0.4844905070216916], "95": [0.4327500906125407, 0.49103704697586353]}, "precision": 0.4790652385589094, "recall": 0.44767970882620567}, "group": {"f1": 0.593177511054959, "f1_ci": {"90": [0.5717044673748471, 0.6144093322442311], "95": [0.5676831336909995, 0.6175847266352983]}, "precision": 0.5697815533980582, "recall": 0.6185770750988142}, "location": {"f1": 0.6333949476278496, "f1_ci": {"90": [0.6061374762309342, 0.6596729625322199], "95": [0.5999818181818181, 0.6642491807610421]}, "precision": 0.5667034178610805, "recall": 0.7178770949720671}, "person": {"f1": 0.8279457768508863, "f1_ci": {"90": [0.8177849742494966, 0.8392861816686192], "95": [0.8161339468587502, 0.8409499904355973]}, "precision": 0.7830374753451677, "recall": 0.8783185840707964}, "product": {"f1": 0.631, "f1_ci": {"90": [0.6078329186866498, 0.6538099694036692], "95": [0.602711796196466, 0.6565176505572475]}, "precision": 0.6138132295719845, "recall": 0.6491769547325102}}}
eval/metric_span.test_2020.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"micro/f1": 0.7595141700404859, "micro/f1_ci": {}, "micro/recall": 0.7301504929942917, "micro/precision": 0.7913385826771654, "macro/f1": 0.7595141700404859, "macro/f1_ci": {}, "macro/recall": 0.7301504929942917, "macro/precision": 0.7913385826771654}
eval/metric_span.test_2021.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"micro/f1": 0.7788214245778822, "micro/f1_ci": {}, "micro/recall": 0.8054816699433329, "micro/precision": 0.7538694663924668, "macro/f1": 0.7788214245778822, "macro/f1_ci": {}, "macro/recall": 0.8054816699433329, "macro/precision": 0.7538694663924668}
eval/prediction.2020.test.json DELETED
The diff for this file is too large to render. See raw diff
 
eval/prediction.2021.dev.json DELETED
The diff for this file is too large to render. See raw diff
 
eval/prediction.2021.test.json DELETED
The diff for this file is too large to render. See raw diff
 
trainer_config.json CHANGED
@@ -1 +1 @@
1
- {"data_split": "2021.train", "model": "vinai/bertweet-base", "crf": false, "max_length": 128, "epoch": 30, "batch_size": 32, "lr": 0.0001, "random_seed": 0, "gradient_accumulation_steps": 1, "weight_decay": 1e-07, "lr_warmup_step_ratio": 0.3, "max_grad_norm": 1}
 
1
+ {"dataset": ["tner/tweetner7"], "dataset_split": "train_2021", "dataset_name": null, "local_dataset": null, "model": "vinai/bertweet-base", "crf": false, "max_length": 128, "epoch": 30, "batch_size": 32, "lr": 0.0001, "random_seed": 0, "gradient_accumulation_steps": 1, "weight_decay": 1e-07, "lr_warmup_step_ratio": 0.3, "max_grad_norm": 1}