blob: cf75522853da8aa1ebd7e6121945a555428e6b4f (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
|
# Evaluating cross-lingual textual similarity on dictionary alignment
This repository contains the scripts to prepare the resources for the study as well as open source implementations of the methods.
## Requirements
- python3
- nltk
```python
import nltk
nltk.download('wordnet')
```
## Acquiring The Data
```bash
git clone https://github.com/yigitsever/Evaluating-Dictionary-Alignment.git && cd Evaluating-Dictionary-Alignment
./get_data.sh
```
This will create two directories; `dictionaries` and `wordnets`.
Linewise aligned definition files are in `wordnets/ready`.
## Acquiring The Embeddings
We use [VecMap](https://github.com/artetxem/vecmap) on [fastText](https://fasttext.cc/) embeddings.
You can skip this step if you are providing your own polylingual embeddings.
Otherwise simply run;
```bash
./get_embeddings.sh
```
Bear in mind that this will require around 30 GB free space.
|
