Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/MilaNLProc/bertlang
A web interface to understand language-specific BERT-models
https://github.com/MilaNLProc/bertlang
artificial-intelligence bert-model machine-learning nlp nlp-machine-learning
Last synced: 23 days ago
JSON representation
A web interface to understand language-specific BERT-models
- Host: GitHub
- URL: https://github.com/MilaNLProc/bertlang
- Owner: MilaNLProc
- License: mit
- Created: 2020-02-20T08:38:42.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2024-04-16T21:46:30.000Z (5 months ago)
- Last Synced: 2024-06-22T15:54:34.053Z (3 months ago)
- Topics: artificial-intelligence, bert-model, machine-learning, nlp, nlp-machine-learning
- Language: JavaScript
- Homepage: http://bertlang.unibocconi.it
- Size: 4.02 MB
- Stars: 17
- Watchers: 4
- Forks: 4
- Open Issues: 6
-
Metadata Files:
- Readme: README.md
- License: LICENSE
- Security: SECURITY.md
Awesome Lists containing this project
README
# [BertLang](https://bertlang.unibocconi.it)
[BertLang](https://bertlang.unibocconi.it) is a webapp that contains info about language-specific BERT models.
## How to Contribute
This is a **collaborative** resource to help researchers understand and find the best BERT model for a given dataset, task and language. The numbers here rely on self reported performance (we can give no guarantees for their accuracy. In the future, we hope to independently verify each of the models).
We currently store all the information in a .json file `static/data/data_example.json`. We are keeping this structure that is easy to parse and to check.
**Do you want to add a new model or suggest updates?** Send us a pull request! Please note that we aim for consistency in the performance metric across tasks (e.g. Sentiment Analysis -> Accuracy).
See the following example for the Italian BERT model, ALBERTO.
```json
{
"name": "ALBERTO",
"language": "Italian",
"tasks": [
{
"source": "http://ceur-ws.org/Vol-2481/paper57.pdf",
"code": "https://github.com/marcopoli/AlBERTo-it",
"name": "SA",
"dataset": {
"name": "SENTIPOLC 2016",
"link": "http://www.di.unito.it/~tutreeb/sentipolc-evalita16/data.html",
"domain": "twitter"
},
"measure": "F1 (test)",
"performance": 72.23,
"multi_lingual": "nan",
"multi_difference": "nan"
},
{
"name": "SC",
"source": "http://ceur-ws.org/Vol-2481/paper57.pdf",
"code": "https://github.com/marcopoli/AlBERTo-it",
"dataset": {
"name": "SENTIPOLC 2016",
"link": "http://www.di.unito.it/~tutreeb/sentipolc-evalita16/data.html",
"domain": "twitter"
},
"measure": "F1 (test)",
"performance": 79.06,
"multi_lingual": "nan",
"multi_difference": "nan"
},
{
"name": "ID",
"source": "http://ceur-ws.org/Vol-2481/paper57.pdf",
"code": "https://github.com/marcopoli/AlBERTo-it",
"dataset": {
"name": "SENTIPOLC 2016",
"link": "http://www.di.unito.it/~tutreeb/sentipolc-evalita16/data.html",
"domain": "twitter"
},
"measure": "F1 (test)",
"performance": 60.9,
"multi_lingual": "nan",
"multi_difference": "nan"
}
]
}
```
## NLP Task Acronyms
Please refer to this table for using the correct NLP task acronym.
| NLP task | Acronym |
|--------------------|-----------------------------------|
| POS | Part of Speech Tagging |
| DP | Dependency Parsing |
| NER | Named Entity Recognition |
| NLI | Natural Language Inference |
| PI | Paraphrase Identification |
| STS | Semantic Textual Similarity |
| WSD | Word Sense Disambiguation |
| TC | Text Classification |
| CP | Constituency Parsing |
| SA | Sentiment Analysis |
| SRL | Semantic Role Labeling |
| STR | Spatio-Temporal Relation |
| LPR | Linguistic Properties Recognition |
| OLI | Offensive Language Identification |
| DP-UAS | Unlabeled Attachment Score |
| DP-LAS | Labeled Attachment Score |
| VSD | Verb Sense Disambiguation |
| NSD | Noun Sense Disambiguation |
| SC | Subjectivity Classification |
| ID | Irony Detection |
| DDD | Die/Dat Disambiguation |
| MRC | Machine Reading Comprehension |
| SPM | Sentence Pair Matching |
| POS (coarse) | Part of Speech Tagging |
| POS (fine-grained) | Part of Speech Tagging |
| XPOS | Language-specific POS tagging |
| Morph | Morphological tagging |
| LA | Linguistic Acceptability |
| TER | Textual Entailment Recognition |
| QA | Question Answering |
| CI | Commonsense Inference |
| RC | Reading Comprehension |
## Contributors
+ Debora Nozza - [Twitter](http://twitter.com/debora_nozza) | [Personal Website](http://dnozza.github.io/) | [email protected]
+ Federico Bianchi - [Twitter](http://twitter.com/fb_vinid) | [Personal Website](http://vinid.io/) | [email protected]
+ Dirk Hovy - [Twitter](http://twitter.com/dirk_hovy) | [Personal Website](http://dirkhovy.com/) | [email protected]
## Copyright and License
Built with [Start Bootstrap](https://startbootstrap.com/template-overviews/bare/).
Start Bootstrap is an open source library of free Bootstrap templates and themes. All of the free templates and themes on Start Bootstrap are released under the MIT license, which means you can use them for any purpose, even for commercial projects.
Copyright 2013-2019 Blackrock Digital LLC. Code released under the [MIT](https://github.com/BlackrockDigital/startbootstrap-bare/blob/gh-pages/LICENSE) license.