Translation
Fairseq
AlexTheSun commited on
Commit
2d760d8
1 Parent(s): 01eadc0

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +26 -26
README.md CHANGED
@@ -8,31 +8,6 @@ The TartuNLP Multilingual Neural Machine Translation model for low-resource Finn
8
  - **High and Mid-Resource Languages:** Estonian, English, Finnish, Hungarian, Latvian, Norwegian, Russian
9
  - **Low-Resource Finno-Ugric Languages:** Komi, Komi Permyak, Udmurt, Hill Mari, Meadow Mari, Erzya, Moksha, Proper Karelian, Livvi Karelian, Ludian, Võro, Veps, Livonian, Northern Sami, Southern Sami, Inari Sami, Lule Sami, Skolt Sami, Mansi, Khanty
10
 
11
- ### Scores
12
- Average:
13
- | to-lang | bleu | chrf | chrf++ |
14
- | ------- | ----- | ---- | ------ |
15
- | ru | 24.82 | 51.81 | 49.08 |
16
- | en | 28.24 | 55.91 | 53.73 |
17
- | et | 18.66 | 51.72 | 47.69 |
18
- | fi | 15.45 | 50.04 | 45.38 |
19
- | hun | 16.73 | 47.38 | 44.19 |
20
- | lv | 18.15 | 49.04 | 45.54 |
21
- | nob | 14.43 | 45.64 | 42.29 |
22
- | kpv | 10.73 | 42.34 | 38.50 |
23
- | liv | 5.16 | 29.95 | 27.28 |
24
- | mdf | 5.27 | 37.66 | 32.99 |
25
- | mhr | 8.51 | 43.42 | 38.76 |
26
- | mns | 2.45 | 27.75 | 24.03 |
27
- | mrj | 7.30 | 40.81 | 36.40 |
28
- | myv | 4.72 | 38.74 | 33.80 |
29
- | olo | 4.63 | 34.43 | 30.00 |
30
- | udm | 7.50 | 40.07 | 35.72 |
31
- | krl | 9.39 | 42.74 | 38.24 |
32
- | vro | 8.64 | 39.89 | 35.97 |
33
- | vep | 6.73 | 38.15 | 33.91 |
34
- | lud | 3.11 | 31.50 | 27.30 |
35
-
36
  ### Usage
37
  To use this model for translation tasks, you will need to utilize the [**Fairseq v0.12.2**](https://pypi.org/project/fairseq/0.12.2/).
38
 
@@ -67,4 +42,29 @@ fairseq-interactive ${model_path} \
67
  --lang-pairs ${src_lang}-${tgt_lang} \
68
  --langs "${nllb_langs},${new_langs}" \
69
  --cpu
70
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8
  - **High and Mid-Resource Languages:** Estonian, English, Finnish, Hungarian, Latvian, Norwegian, Russian
9
  - **Low-Resource Finno-Ugric Languages:** Komi, Komi Permyak, Udmurt, Hill Mari, Meadow Mari, Erzya, Moksha, Proper Karelian, Livvi Karelian, Ludian, Võro, Veps, Livonian, Northern Sami, Southern Sami, Inari Sami, Lule Sami, Skolt Sami, Mansi, Khanty
10
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
11
  ### Usage
12
  To use this model for translation tasks, you will need to utilize the [**Fairseq v0.12.2**](https://pypi.org/project/fairseq/0.12.2/).
13
 
 
42
  --lang-pairs ${src_lang}-${tgt_lang} \
43
  --langs "${nllb_langs},${new_langs}" \
44
  --cpu
45
+ ```
46
+
47
+ ### Scores
48
+ Average:
49
+ | to-lang | bleu | chrf | chrf++ |
50
+ | ------- | ----- | ---- | ------ |
51
+ | ru | 24.82 | 51.81 | 49.08 |
52
+ | en | 28.24 | 55.91 | 53.73 |
53
+ | et | 18.66 | 51.72 | 47.69 |
54
+ | fi | 15.45 | 50.04 | 45.38 |
55
+ | hun | 16.73 | 47.38 | 44.19 |
56
+ | lv | 18.15 | 49.04 | 45.54 |
57
+ | nob | 14.43 | 45.64 | 42.29 |
58
+ | kpv | 10.73 | 42.34 | 38.50 |
59
+ | liv | 5.16 | 29.95 | 27.28 |
60
+ | mdf | 5.27 | 37.66 | 32.99 |
61
+ | mhr | 8.51 | 43.42 | 38.76 |
62
+ | mns | 2.45 | 27.75 | 24.03 |
63
+ | mrj | 7.30 | 40.81 | 36.40 |
64
+ | myv | 4.72 | 38.74 | 33.80 |
65
+ | olo | 4.63 | 34.43 | 30.00 |
66
+ | udm | 7.50 | 40.07 | 35.72 |
67
+ | krl | 9.39 | 42.74 | 38.24 |
68
+ | vro | 8.64 | 39.89 | 35.97 |
69
+ | vep | 6.73 | 38.15 | 33.91 |
70
+ | lud | 3.11 | 31.50 | 27.30 |