Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- el
|
4 |
+
---
|
5 |
+
|
6 |
+
# PaloBERT
|
7 |
+
|
8 |
+
A greek pre-trained language model based on [RoBERTa](https://arxiv.org/abs/1907.11692).
|
9 |
+
|
10 |
+
## Pre-training data
|
11 |
+
|
12 |
+
The model is pre-trained on a corpus of 458,293 documents collected from greek social media (Twitter, Instagram, Facebook and YouTube). A RoBERTa tokenizer trained from scratch on the same corpus is also included.
|
13 |
+
|
14 |
+
The corpus has been provided by [Palo LTD](http://www.paloservices.com/)
|
15 |
+
|
16 |
+
|
17 |
+
## Requirements
|
18 |
+
|
19 |
+
```
|
20 |
+
pip install transformers
|
21 |
+
pip install torch
|
22 |
+
|
23 |
+
```
|
24 |
+
|
25 |
+
## Pre-processing details
|
26 |
+
|
27 |
+
In order to use 'palobert-base-greek-social-media', the text needs to be pre-processed as follows:
|
28 |
+
|
29 |
+
* remove all greek diacritics
|
30 |
+
* convert to lowercase
|
31 |
+
* remove all punctuation
|
32 |
+
|
33 |
+
|
34 |
+
## Evaluation on MLM and Sentiment Analysis tasks
|
35 |
+
|
36 |
+
For detailed results refer to Thesis: ['Ανάλυση συναισθήματος κειμένου στα Ελληνικά με χρήση Δικτύων Μετασχηματιστών']( http://artemis.cslab.ece.ntua.gr:8080/jspui/handle/123456789/18623) (version - p2)
|
37 |
+
|
38 |
+
## Author
|
39 |
+
|
40 |
+
Pavlina Chatziantoniou, Georgios Alexandridis and Athanasios Voulodimos
|
41 |
+
|
42 |
+
## Citation info
|
43 |
+
|
44 |
+
http://artemis.cslab.ece.ntua.gr:8080/jspui/handle/123456789/18623
|