Commit
·
8aa487b
1
Parent(s):
a2ea80f
Update README.md
Browse files
README.md
CHANGED
|
@@ -7,14 +7,22 @@ language:
|
|
| 7 |
metrics:
|
| 8 |
- perplexity
|
| 9 |
library_name: transformers
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
---
|
| 11 |
|
| 12 |
## Twitter-roBERTa-base fine-tuned using masked language modelling
|
| 13 |
-
This is a RoBERTa-base model finetuned (domain adaptation) on ~2M tweets from Jin 2009.
|
| 14 |
This is the first step of a two steps approach to finetune for sentiment analysis (ULMFit)
|
| 15 |
This model is suitable for English.
|
| 16 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
Reference Paper: [ULMFit](https://arxiv.org/abs/1801.06146).
|
|
|
|
| 18 |
Git Repo: TBD
|
| 19 |
-
Labels: 0 -> Negative; 1 -> Positive
|
| 20 |
-
|
|
|
|
| 7 |
metrics:
|
| 8 |
- perplexity
|
| 9 |
library_name: transformers
|
| 10 |
+
tags:
|
| 11 |
+
- distillroberta-base
|
| 12 |
+
- twitter
|
| 13 |
+
pipeline_tag: fill-mask
|
| 14 |
---
|
| 15 |
|
| 16 |
## Twitter-roBERTa-base fine-tuned using masked language modelling
|
| 17 |
+
This is a RoBERTa-base model finetuned (domain adaptation) on ~2M tweets from Jin 2009 (sentiment140).
|
| 18 |
This is the first step of a two steps approach to finetune for sentiment analysis (ULMFit)
|
| 19 |
This model is suitable for English.
|
| 20 |
|
| 21 |
+
Main charachetistics:
|
| 22 |
+
- pretrained model and tokenizer: distillroberta-base
|
| 23 |
+
- no cleaning/processing applied to the data
|
| 24 |
+
|
| 25 |
Reference Paper: [ULMFit](https://arxiv.org/abs/1801.06146).
|
| 26 |
+
Reference dataset: [Sentiment140](https://www.kaggle.com/datasets/kazanova/sentiment140?resource=download)
|
| 27 |
Git Repo: TBD
|
| 28 |
+
Labels: 0 -> Negative; 1 -> Positive
|
|
|