Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- word2vec
|
4 |
+
language: de
|
5 |
+
license: mit
|
6 |
+
datasets:
|
7 |
+
- wikipedia
|
8 |
+
---
|
9 |
+
|
10 |
+
## Description
|
11 |
+
German word embedding model trained by Müller with the following parameter configuration:
|
12 |
+
- a corpus as big as possible (and as diverse as possible without being informal) filtering of punctuation and stopwords
|
13 |
+
- forming bigramm tokens
|
14 |
+
- using skip-gram as training algorithm with hierarchical softmax
|
15 |
+
- window size between 5 and 10
|
16 |
+
- dimensionality of feature vectors of 300 or more
|
17 |
+
- using negative sampling with 10 samples
|
18 |
+
- ignoring all words with total frequency lower than 50
|
19 |
+
|
20 |
+
For more information, see [https://devmount.github.io/GermanWordEmbeddings/](https://devmount.github.io/GermanWordEmbeddings/)
|
21 |
+
|
22 |
+
## How to use?
|
23 |
+
|
24 |
+
```
|
25 |
+
from gensim.models import KeyedVectors
|
26 |
+
from huggingface_hub import hf_hub_download
|
27 |
+
model = KeyedVectors.load_word2vec_format(hf_hub_download(repo_id="Word2vec/german_model", filename="german.model"), binary=True, unicode_errors="ignore")
|
28 |
+
model.most_similar("exemple")```
|
29 |
+
|
30 |
+
## Citation
|
31 |
+
|
32 |
+
```
|
33 |
+
@thesis{mueller2015,
|
34 |
+
author = {{Müller}, Andreas},
|
35 |
+
title = "{Analyse von Wort-Vektoren deutscher Textkorpora}",
|
36 |
+
school = {Technische Universität Berlin},
|
37 |
+
year = 2015,
|
38 |
+
month = jun,
|
39 |
+
type = {Bachelor's Thesis},
|
40 |
+
url = {https://devmount.github.io/GermanWordEmbeddings}
|
41 |
+
}
|
42 |
+
```
|