Translation
Fairseq
English
Icelandic
wmt
atlijas commited on
Commit
318d76a
·
verified ·
1 Parent(s): 4e3218c

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +54 -3
README.md CHANGED
@@ -1,3 +1,54 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ language:
4
+ - en
5
+ - is
6
+ library_name: fairseq
7
+ tags:
8
+ - translation
9
+ - wmt
10
+ ---
11
+
12
+ ## Model description
13
+ This is a translation model which translates text from English to Icelandic. It follows the architecture of the transformer model described in [Attention is All You Need](https://arxiv.org/pdf/1706.03762) and was trained with [fairseq](https://github.com/facebookresearch/fairseq) for [WMT24](https://www2.statmt.org/wmt24/).
14
+
15
+ This is the base version of our model. See also: [wmt24-en-is-transformer-base](https://huggingface.co/arnastofnun/wmt24-en-is-transformer-base), [wmt24-en-is-transformer-big](https://huggingface.co/arnastofnun/wmt24-en-is-transformer-base-deep), [wmt24-en-is-transformer-big-deep](https://huggingface.co/arnastofnun/wmt24-en-is-transformer-big-deep).
16
+
17
+ | model | d_model | d_ff | h | N_enc | N_dec |
18
+ |:---------------|:----------------------|:-------------------|:--------------|:--------------------|:--------------------|
19
+ | Base | 512 | 2048 | 8 | 6 | 6 |
20
+ | Base_deep | 512 | 2048 | 8 | 36 | 12 |
21
+ | Big | 1024 | 4096 | 16 | 6 | 6 |
22
+ | Big_deep | 1024 | 4096 | 16 | 36 | 12 |
23
+
24
+
25
+ #### How to use
26
+
27
+ ```python
28
+ from fairseq.models.transformer import TransformerModel
29
+ TRANSLATION_MODEL_NAME = 'checkpoint_best.pt'
30
+ TRANSLATION_MODEL = TransformerModel.from_pretrained('path/to/model', checkpoint_file=TRANSLATION_MODEL_NAME, bpe='sentencepiece', sentencepiece_model='sentencepiece.bpe.model')
31
+ src_sentences = ['This is a test sentence.', 'This is another test sentence.']
32
+ translated_sentences = translate(translation_model=TRANSLATION_MODEL, sentences=src_sentences, beam=5)
33
+ print(translated_sentences)
34
+ ```
35
+
36
+ ## Eval results
37
+ We evaluated our data on the [WMT21 test set](https://github.com/wmt-conference/wmt21-news-systems/). These are the chrF scores for our published models:
38
+
39
+ | model | chrF |
40
+ |:---------------|:------|
41
+ | Base | 56.8 |
42
+ | Base_deep | 57.1 |
43
+ | Big | 57.7 |
44
+ | Big_deep | 57.7 |
45
+ ## BibTeX entry and citation info
46
+
47
+ ```bibtex
48
+ @inproceedings{...,
49
+ year={XXX},
50
+ title={XXX},
51
+ author={XXX},
52
+ booktitle={XXX},
53
+ }
54
+ ```