Commit
•
893342d
1
Parent(s):
b4b3edc
update README
Browse files
README.md
CHANGED
@@ -1,24 +1,24 @@
|
|
1 |
---
|
2 |
language:
|
3 |
- en
|
4 |
-
-
|
5 |
tags:
|
6 |
- translation
|
7 |
license: cc-by-4.0
|
8 |
---
|
9 |
|
10 |
-
|
11 |
|
12 |
-
This repository contains the translation model for en-
|
13 |
|
14 |
### Model Info
|
15 |
|
16 |
-
* Source language:
|
17 |
-
* Target language:
|
18 |
-
*
|
19 |
* Model architecture: Transformer-base
|
20 |
* Tokenizer: SentencePiece (Unigram)
|
21 |
-
* Cleaning: We used OpusCleaner with a set of basic rules. Details can be found in the filter files in [Github](https://github.com/hplt-project/HPLT-MT-Models/tree/main/v1.0/data/en-
|
22 |
|
23 |
You can also read our deliverable report [here](https://hplt-project.org/HPLT_D5_1___Translation_models_for_select_language_pairs.pdf) for more details.
|
24 |
|
@@ -29,12 +29,12 @@ The model has been trained with Marian. To run inference, refer to the [Inferenc
|
|
29 |
|
30 |
The model can be used with the Hugging Face framework if the weights are converted to the Hugging Face format. We might provide this in the future; contributions are also welcome.
|
31 |
|
32 |
-
|
33 |
|
34 |
| testset | BLEU | chrF++ | COMET22 |
|
35 |
| -------------------------------------- | ---- | ----- | ----- |
|
36 |
-
| flores200 |
|
37 |
-
| ntrex |
|
38 |
|
39 |
### Acknowledgements
|
40 |
|
|
|
1 |
---
|
2 |
language:
|
3 |
- en
|
4 |
+
- sq
|
5 |
tags:
|
6 |
- translation
|
7 |
license: cc-by-4.0
|
8 |
---
|
9 |
|
10 |
+
## HPLT MT release v1.0
|
11 |
|
12 |
+
This repository contains the translation model for en-sq trained with OPUS and HPLT data. For usage instructions, evaluation scripts, and inference scripts, please refer to the [HPLT-MT-Models v1.0](https://github.com/hplt-project/HPLT-MT-Models/tree/main/v1.0) GitHub repository.
|
13 |
|
14 |
### Model Info
|
15 |
|
16 |
+
* Source language: English
|
17 |
+
* Target language: Albanian
|
18 |
+
* Dataset: OPUS and HPLT data
|
19 |
* Model architecture: Transformer-base
|
20 |
* Tokenizer: SentencePiece (Unigram)
|
21 |
+
* Cleaning: We used OpusCleaner with a set of basic rules. Details can be found in the filter files in [Github](https://github.com/hplt-project/HPLT-MT-Models/tree/main/v1.0/data/en-sq/raw/v2)
|
22 |
|
23 |
You can also read our deliverable report [here](https://hplt-project.org/HPLT_D5_1___Translation_models_for_select_language_pairs.pdf) for more details.
|
24 |
|
|
|
29 |
|
30 |
The model can be used with the Hugging Face framework if the weights are converted to the Hugging Face format. We might provide this in the future; contributions are also welcome.
|
31 |
|
32 |
+
### Benchmarks
|
33 |
|
34 |
| testset | BLEU | chrF++ | COMET22 |
|
35 |
| -------------------------------------- | ---- | ----- | ----- |
|
36 |
+
| flores200 | 30.7 | 56.6 | 0.8761 |
|
37 |
+
| ntrex | 32.7 | 56.1 | 0.8517 |
|
38 |
|
39 |
### Acknowledgements
|
40 |
|