Update README.md
Browse files
README.md
CHANGED
|
@@ -14,18 +14,6 @@ widget:
|
|
| 14 |
|
| 15 |
Fine-tuned model on English language using a masked language modeling (MLM) objective from BERTweet in [this repository](https://github.com/VinAIResearch/BERTweet) for the classification task for false/misleading information about COVID-19 vaccines.
|
| 16 |
|
| 17 |
-
## Model description
|
| 18 |
-
|
| 19 |
-
- Baseline model: BERTweet14,15
|
| 20 |
-
- trained based on the RoBERTa pre-training procedure
|
| 21 |
-
- 850M General English Tweets (Jan 2012 ~ Aug 2019)
|
| 22 |
-
- 23M COVID-19 English Tweets
|
| 23 |
-
- Size of the model: >134M parameters
|
| 24 |
-
- Further training
|
| 25 |
-
- Training with recent COVID-19 and vaccine tweets
|
| 26 |
-
|
| 27 |
-
You can embed local or remote images using ``
|
| 28 |
-
|
| 29 |
## Intended uses & limitations
|
| 30 |
|
| 31 |
#### How to use
|
|
@@ -48,34 +36,30 @@ Provide examples of latent issues and potential remediations.
|
|
| 48 |
#### 2) Fine-tuning for fact classification
|
| 49 |
- Statements from Poynter and Snopes with Selenium 14,000 fact-checked statements from 2020-01-14 to 2021-05-13
|
| 50 |
- Divide original labels within 3 categories
|
| 51 |
-
False:
|
| 52 |
-
Misleading:
|
| 53 |
-
True
|
| 54 |
|
| 55 |
Describe the data you used to train the model.
|
| 56 |
If you initialized it with pre-trained weights, add a link to the pre-trained model card or repository with description of the pre-training data.
|
| 57 |
|
| 58 |
## Training procedure
|
| 59 |
|
| 60 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 61 |
|
| 62 |
## Eval results
|
| 63 |
|
| 64 |
-
### BibTeX entry and citation info
|
| 65 |
-
|
| 66 |
-
```bibtex
|
| 67 |
-
@inproceedings{...,
|
| 68 |
-
year={2020}
|
| 69 |
-
}
|
| 70 |
-
```
|
| 71 |
# Contributors
|
| 72 |
-
-
|
| 73 |
-
-
|
| 74 |
-
- An, Jiyong
|
| 75 |
-
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
- Advisor: Wen-Syan Li
|
| 80 |
-
|
| 81 |
-
|
|
|
|
| 14 |
|
| 15 |
Fine-tuned model on English language using a masked language modeling (MLM) objective from BERTweet in [this repository](https://github.com/VinAIResearch/BERTweet) for the classification task for false/misleading information about COVID-19 vaccines.
|
| 16 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
## Intended uses & limitations
|
| 18 |
|
| 19 |
#### How to use
|
|
|
|
| 36 |
#### 2) Fine-tuning for fact classification
|
| 37 |
- Statements from Poynter and Snopes with Selenium 14,000 fact-checked statements from 2020-01-14 to 2021-05-13
|
| 38 |
- Divide original labels within 3 categories
|
| 39 |
+
False: \\\\\\\\t\\\\\\\\tFalse, no evidence, manipulated, fake, not true, unproven, unverified
|
| 40 |
+
Misleading: \\\\\\\\tMisleading, exaggerated, out of context, needs context
|
| 41 |
+
True:\\\\\\\\t\\\\\\\\tTrue, correct
|
| 42 |
|
| 43 |
Describe the data you used to train the model.
|
| 44 |
If you initialized it with pre-trained weights, add a link to the pre-trained model card or repository with description of the pre-training data.
|
| 45 |
|
| 46 |
## Training procedure
|
| 47 |
|
| 48 |
+
- Baseline model: [BERTweet](https://github.com/VinAIResearch/BERTweet)
|
| 49 |
+
- trained based on the RoBERTa pre-training procedure
|
| 50 |
+
- 850M General English Tweets (Jan 2012 ~ Aug 2019)
|
| 51 |
+
- 23M COVID-19 English Tweets
|
| 52 |
+
- Size of the model: >134M parameters
|
| 53 |
+
- Further training
|
| 54 |
+
- Training with recent COVID-19 and vaccine tweets
|
| 55 |
|
| 56 |
## Eval results
|
| 57 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
# Contributors
|
| 59 |
+
- This page is a part of final team project from MLDL for DS class at SNU
|
| 60 |
+
- Team BIBI - Vaccinating COVID-NineTweets
|
| 61 |
+
- Team members: Ahn, Hyunju; An, Jiyong; An, Seungchan; Jeong, Seokho; Kim, Jungmin; Kim, Sangbeom
|
| 62 |
+
- Advisor: Prof. Wen-Syan Li
|
| 63 |
+
|
| 64 |
+
# 
|
| 65 |
+
<img src="https://gsds.snu.ac.kr/sites/gsds.snu.ac.kr/files/GSDS_logo.png" width="100" height="100">
|
|
|
|
|
|
|
|
|