jramompichel commited on
Commit
9b024b5
1 Parent(s): 42647dd

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +13 -12
README.md CHANGED
@@ -5,6 +5,7 @@ license: mit
5
  **Descrición do Modelo / Model description**
6
 
7
  Modelo feito con OpenNMT para o par español-galego utilizando unha arquitectura transformer.
 
8
  Model developed with OpenNMT for the Spanish-Galician pair using a transformer architecture.
9
 
10
  **Como utilizar**
@@ -22,44 +23,44 @@ onmt_translate -src input_text -model NOS-MT-es-gl -output ./output_file.txt -r
22
  **Adestramento / Training**
23
 
24
  No adestramento, utilizamos corpora auténticos e sintéticos. Os primeiros son corpora de traducións feitas directamente por tradutores humanos. Os segundos son corpora de traducións español-portugués e inglés-portugués, que convertemos en español-galego e inglés-galego a través da tradución automática portugués-galego con Opentrad/Apertium e transliteración para palabras fóra de vocabulário.
 
25
  In the training we have used authentic and synthetic corpora. The former are corpora of translations directly produced by human translators. The latter are corpora of Spanish-Portuguese and English-Portuguese translations, which we have converted into Spanish-Galician and English-Galician by means of Portuguese-Galician translation with Opentrad/Apertium and transliteration for out-of-vocabulary words.
26
 
27
 
28
- **Procedemento de adestramento**
29
 
30
- + Tokenization dos datasets feita co tokenizador de linguakit https://github.com/citiususc/Linguakit
31
 
32
- + O vocabulario para os modelos foi xerado a través do script [learn_bpe.py](https://github.com/OpenNMT/OpenNMT-py/blob/master/tools/learn_bpe.py) da open NMT
33
 
34
- + Usando o .yaml neste repositorio pode replicar o proceso de adestramento do seguinte xeito
35
 
36
  ```bash
37
  onmt_build_vocab -config bpe-es-gl_emb.yaml -n_sample 100000
38
  onmt_train -config bpe-es-gl_emb.yaml
39
  ```
40
 
41
- **Hiperparámetros**
42
 
43
  Os parámetros usados para o desenvolvimento do modelo poden ser consultados directamente no mesmo ficheiro .yaml bpe-es-gl_emb.yaml
44
 
 
 
 
45
 
46
- **Avaliación**
47
- A avalación dos modelos é feita cunha mistura de tests desenvolvidos internamente
48
- (gold1, gold2, test-suite) con outros datasets disponíbeis en galego (Flores).
49
 
50
  | GOLD 1 | GOLD 2 | FLORES | TEST-SUITE|
51
  | ------------- |:-------------:| -------:|----------:|
52
  | 79.6 | 43.3 | 21.8 | 74.3 |
53
 
54
-
55
-
56
  **Información adicional**
57
 
58
  Licensing information
59
 
60
- Apache License, Version 2.0
61
 
62
- **Financiamento**
63
 
64
  This research was funded by the project "Nós: Galician in the society and economy of artificial intelligence", agreement between Xunta de Galicia and University of Santiago de Compostela, and grant ED431G2019/04 by the Galician Ministry of Education, University and Professional Training, and the European Regional Development Fund (ERDF/FEDER program), and Groups of Reference: ED431C 2020/21.
65
 
 
5
  **Descrición do Modelo / Model description**
6
 
7
  Modelo feito con OpenNMT para o par español-galego utilizando unha arquitectura transformer.
8
+
9
  Model developed with OpenNMT for the Spanish-Galician pair using a transformer architecture.
10
 
11
  **Como utilizar**
 
23
  **Adestramento / Training**
24
 
25
  No adestramento, utilizamos corpora auténticos e sintéticos. Os primeiros son corpora de traducións feitas directamente por tradutores humanos. Os segundos son corpora de traducións español-portugués e inglés-portugués, que convertemos en español-galego e inglés-galego a través da tradución automática portugués-galego con Opentrad/Apertium e transliteración para palabras fóra de vocabulário.
26
+
27
  In the training we have used authentic and synthetic corpora. The former are corpora of translations directly produced by human translators. The latter are corpora of Spanish-Portuguese and English-Portuguese translations, which we have converted into Spanish-Galician and English-Galician by means of Portuguese-Galician translation with Opentrad/Apertium and transliteration for out-of-vocabulary words.
28
 
29
 
30
+ **Procedemento de adestramento / Training process**
31
 
32
+ + Tokenization dos datasets feita co tokenizador de linguakit / Tokenization of the datasets made with linguakit tokeniser https://github.com/citiususc/Linguakit
33
 
34
+ + O vocabulario para os modelos foi xerado a través do script / Vocabulary for the models was created by the script [learn_bpe.py](https://github.com/OpenNMT/OpenNMT-py/blob/master/tools/learn_bpe.py) da open NMT
35
 
36
+ + Usando o .yaml neste repositorio pode replicar o proceso de adestramento do seguinte xeito / Using the .yaml in this repository you can replicate the training process as follows
37
 
38
  ```bash
39
  onmt_build_vocab -config bpe-es-gl_emb.yaml -n_sample 100000
40
  onmt_train -config bpe-es-gl_emb.yaml
41
  ```
42
 
43
+ **Hiperparámetros / Hyper-parameters**
44
 
45
  Os parámetros usados para o desenvolvimento do modelo poden ser consultados directamente no mesmo ficheiro .yaml bpe-es-gl_emb.yaml
46
 
47
+ The parameters used for the development of the model can be directly viewed in the same .yaml file bpe-es-gl_emb.yaml
48
+
49
+ **Avaliación / Evaluation**
50
 
51
+ A avalación dos modelos é feita cunha mistura de tests desenvolvidos internamente (gold1, gold2, test-suite) con outros datasets disponíbeis en galego (Flores).
 
 
52
 
53
  | GOLD 1 | GOLD 2 | FLORES | TEST-SUITE|
54
  | ------------- |:-------------:| -------:|----------:|
55
  | 79.6 | 43.3 | 21.8 | 74.3 |
56
 
 
 
57
  **Información adicional**
58
 
59
  Licensing information
60
 
61
+ MIT
62
 
63
+ **Financiamento / Funding**
64
 
65
  This research was funded by the project "Nós: Galician in the society and economy of artificial intelligence", agreement between Xunta de Galicia and University of Santiago de Compostela, and grant ED431G2019/04 by the Galician Ministry of Education, University and Professional Training, and the European Regional Development Fund (ERDF/FEDER program), and Groups of Reference: ED431C 2020/21.
66