metadata
datasets:
- tner/tweetner7
metrics:
- f1
- precision
- recall
pipeline_tag: token-classification
widget:
- text: >-
Get the all-analog Classic Vinyl Edition of `Takin' Off` Album from
{@herbiehancock@} via {@bluenoterecords@} link below: {{URL}}
example_title: NER Example 1
base_model: roberta-large
model-index:
- name: tner/roberta-large-tweetner7-2020-selflabel2021-all
results:
- task:
type: token-classification
name: Token Classification
dataset:
name: tner/tweetner7
type: tner/tweetner7
args: tner/tweetner7
metrics:
- type: f1
value: 0.6451758087201125
name: F1 (test_2021)
- type: precision
value: 0.6282458639202366
name: Precision (test_2021)
- type: recall
value: 0.6630434782608695
name: Recall (test_2021)
- type: f1_macro
value: 0.5945137835095485
name: Macro F1 (test_2021)
- type: precision_macro
value: 0.5791991181065553
name: Macro Precision (test_2021)
- type: recall_macro
value: 0.6195808065595296
name: Macro Recall (test_2021)
- type: f1_entity_span
value: 0.7849668054461573
name: Entity Span F1 (test_2021)
- type: precision_entity_span
value: 0.7643256272597787
name: Entity Span Precision (test_2020)
- type: recall_entity_span
value: 0.8067537874407309
name: Entity Span Recall (test_2021)
- type: f1
value: 0.6605206073752712
name: F1 (test_2020)
- type: precision
value: 0.6916524701873935
name: Precision (test_2020)
- type: recall
value: 0.6320705760249092
name: Recall (test_2020)
- type: f1_macro
value: 0.6182768841282975
name: Macro F1 (test_2020)
- type: precision_macro
value: 0.646958757311601
name: Macro Precision (test_2020)
- type: recall_macro
value: 0.600022393469146
name: Macro Recall (test_2020)
- type: f1_entity_span
value: 0.769397721106891
name: Entity Span F1 (test_2020)
- type: precision_entity_span
value: 0.8061398521887436
name: Entity Span Precision (test_2020)
- type: recall_entity_span
value: 0.7358588479501816
name: Entity Span Recall (test_2020)
tner/roberta-large-tweetner7-2020-selflabel2021-all
This model is a fine-tuned version of roberta-large on the
tner/tweetner7 dataset (train split). This model is fine-tuned on self-labeled dataset which is the extra_2021 split of the tner/tweetner7 annotated by tner/roberta-large). Please check https://github.com/asahi417/tner/tree/master/examples/tweetner7_paper#model-fine-tuning-self-labeling for more detail of reproducing the model.
Model fine-tuning is done via T-NER's hyper-parameter search (see the repository
for more detail). It achieves the following results on the test set of 2021:
- F1 (micro): 0.6451758087201125
- Precision (micro): 0.6282458639202366
- Recall (micro): 0.6630434782608695
- F1 (macro): 0.5945137835095485
- Precision (macro): 0.5791991181065553
- Recall (macro): 0.6195808065595296
The per-entity breakdown of the F1 score on the test set are below:
- corporation: 0.5067218200620476
- creative_work: 0.45376220562894887
- event: 0.4452749599572877
- group: 0.6063348416289593
- location: 0.6619263089851325
- person: 0.835890955046037
- product: 0.651685393258427
For F1 scores, the confidence interval is obtained by bootstrap as below:
- F1 (micro):
- 90%: [0.6360452531843157, 0.6546242674951402]
- 95%: [0.6344128889037165, 0.6562435584441533]
- F1 (macro):
- 90%: [0.6360452531843157, 0.6546242674951402]
- 95%: [0.6344128889037165, 0.6562435584441533]
Full evaluation can be found at metric file of NER and metric file of entity span.
Usage
This model can be used through the tner library. Install the library via pip.
pip install tner
TweetNER7 pre-processed tweets where the account name and URLs are converted into special formats (see the dataset page for more detail), so we process tweets accordingly and then run the model prediction as below.
import re
from urlextract import URLExtract
from tner import TransformersNER
extractor = URLExtract()
def format_tweet(tweet):
# mask web urls
urls = extractor.find_urls(tweet)
for url in urls:
tweet = tweet.replace(url, "{{URL}}")
# format twitter account
tweet = re.sub(r"\b(\s*)(@[\S]+)\b", r'\1{\2@}', tweet)
return tweet
text = "Get the all-analog Classic Vinyl Edition of `Takin' Off` Album from @herbiehancock via @bluenoterecords link below: http://bluenote.lnk.to/AlbumOfTheWeek"
text_format = format_tweet(text)
model = TransformersNER("tner/roberta-large-tweetner7-2020-selflabel2021-all")
model.predict([text_format])
It can be used via transformers library but it is not recommended as CRF layer is not supported at the moment.
Training hyperparameters
The following hyperparameters were used during training:
- dataset: ['tner/tweetner7']
- dataset_split: train
- dataset_name: None
- local_dataset: {'train': 'tweet_ner/2020_2021.extra.tner/roberta-large-2020.txt', 'validation': 'tweet_ner/2020.dev.txt'}
- model: roberta-large
- crf: True
- max_length: 128
- epoch: 30
- batch_size: 32
- lr: 1e-05
- random_seed: 0
- gradient_accumulation_steps: 1
- weight_decay: 1e-07
- lr_warmup_step_ratio: 0.3
- max_grad_norm: 1
The full configuration can be found at fine-tuning parameter file.
Reference
If you use the model, please cite T-NER paper and TweetNER7 paper.
- T-NER
@inproceedings{ushio-camacho-collados-2021-ner,
title = "{T}-{NER}: An All-Round Python Library for Transformer-based Named Entity Recognition",
author = "Ushio, Asahi and
Camacho-Collados, Jose",
booktitle = "Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations",
month = apr,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.eacl-demos.7",
doi = "10.18653/v1/2021.eacl-demos.7",
pages = "53--62",
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.",
}
- TweetNER7
@inproceedings{ushio-etal-2022-tweet,
title = "{N}amed {E}ntity {R}ecognition in {T}witter: {A} {D}ataset and {A}nalysis on {S}hort-{T}erm {T}emporal {S}hifts",
author = "Ushio, Asahi and
Neves, Leonardo and
Silva, Vitor and
Barbieri, Francesco. and
Camacho-Collados, Jose",
booktitle = "The 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing",
month = nov,
year = "2022",
address = "Online",
publisher = "Association for Computational Linguistics",
}