tiedeman commited on
Commit
84b3a7f
1 Parent(s): 8873ed8

Initial commit

Browse files
.gitattributes CHANGED
@@ -25,3 +25,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
25
  *.zip filter=lfs diff=lfs merge=lfs -text
26
  *.zstandard filter=lfs diff=lfs merge=lfs -text
27
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
25
  *.zip filter=lfs diff=lfs merge=lfs -text
26
  *.zstandard filter=lfs diff=lfs merge=lfs -text
27
  *tfevents* filter=lfs diff=lfs merge=lfs -text
28
+ *.spm filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,171 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - pt
4
+ - ru
5
+ - uk
6
+ - zle
7
+
8
+ tags:
9
+ - translation
10
+
11
+ license: cc-by-4.0
12
+ model-index:
13
+ - name: opus-mt-tc-big-pt-zle
14
+ results:
15
+ - task:
16
+ name: Translation por-rus
17
+ type: translation
18
+ args: por-rus
19
+ dataset:
20
+ name: flores101-devtest
21
+ type: flores_101
22
+ args: por rus devtest
23
+ metrics:
24
+ - name: BLEU
25
+ type: bleu
26
+ value: 26.8
27
+ - task:
28
+ name: Translation por-ukr
29
+ type: translation
30
+ args: por-ukr
31
+ dataset:
32
+ name: flores101-devtest
33
+ type: flores_101
34
+ args: por ukr devtest
35
+ metrics:
36
+ - name: BLEU
37
+ type: bleu
38
+ value: 25.1
39
+ - task:
40
+ name: Translation por-rus
41
+ type: translation
42
+ args: por-rus
43
+ dataset:
44
+ name: tatoeba-test-v2021-08-07
45
+ type: tatoeba_mt
46
+ args: por-rus
47
+ metrics:
48
+ - name: BLEU
49
+ type: bleu
50
+ value: 47.6
51
+ - task:
52
+ name: Translation por-ukr
53
+ type: translation
54
+ args: por-ukr
55
+ dataset:
56
+ name: tatoeba-test-v2021-08-07
57
+ type: tatoeba_mt
58
+ args: por-ukr
59
+ metrics:
60
+ - name: BLEU
61
+ type: bleu
62
+ value: 44.7
63
+ ---
64
+ # opus-mt-tc-big-pt-zle
65
+
66
+ Neural machine translation model for translating from Portuguese (pt) to East Slavic languages (zle).
67
+
68
+ This model is part of the [OPUS-MT project](https://github.com/Helsinki-NLP/Opus-MT), an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of [Marian NMT](https://marian-nmt.github.io/), an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from [OPUS](https://opus.nlpl.eu/) and training pipelines use the procedures of [OPUS-MT-train](https://github.com/Helsinki-NLP/Opus-MT-train).
69
+
70
+ * Publications: [OPUS-MT – Building open translation services for the World](https://aclanthology.org/2020.eamt-1.61/) and [The Tatoeba Translation Challenge – Realistic Data Sets for Low Resource and Multilingual MT](https://aclanthology.org/2020.wmt-1.139/) (Please, cite if you use this model.)
71
+
72
+ ```
73
+ @inproceedings{tiedemann-thottingal-2020-opus,
74
+ title = "{OPUS}-{MT} {--} Building open translation services for the World",
75
+ author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh},
76
+ booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
77
+ month = nov,
78
+ year = "2020",
79
+ address = "Lisboa, Portugal",
80
+ publisher = "European Association for Machine Translation",
81
+ url = "https://aclanthology.org/2020.eamt-1.61",
82
+ pages = "479--480",
83
+ }
84
+
85
+ @inproceedings{tiedemann-2020-tatoeba,
86
+ title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
87
+ author = {Tiedemann, J{\"o}rg},
88
+ booktitle = "Proceedings of the Fifth Conference on Machine Translation",
89
+ month = nov,
90
+ year = "2020",
91
+ address = "Online",
92
+ publisher = "Association for Computational Linguistics",
93
+ url = "https://aclanthology.org/2020.wmt-1.139",
94
+ pages = "1174--1182",
95
+ }
96
+ ```
97
+
98
+ ## Model info
99
+
100
+ * Release: 2022-03-23
101
+ * source language(s): por
102
+ * target language(s): rus ukr
103
+ * valid target language labels: >>rus<< >>ukr<<
104
+ * model: transformer-big
105
+ * data: opusTCv20210807 ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
106
+ * tokenization: SentencePiece (spm32k,spm32k)
107
+ * original model: [opusTCv20210807_transformer-big_2022-03-23.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/por-zle/opusTCv20210807_transformer-big_2022-03-23.zip)
108
+ * more information released models: [OPUS-MT por-zle README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/por-zle/README.md)
109
+ * more information about the model: [MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)
110
+
111
+ This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of `>>id<<` (id = valid target language ID), e.g. `>>rus<<`
112
+
113
+ ## Usage
114
+
115
+ A short example code:
116
+
117
+ ```python
118
+ from transformers import MarianMTModel, MarianTokenizer
119
+
120
+ src_text = [
121
+ ">>ukr<< Esse é o meu lugar.",
122
+ ">>rus<< Tom tem problemas de saúde."
123
+ ]
124
+
125
+ model_name = "pytorch-models/opus-mt-tc-big-pt-zle"
126
+ tokenizer = MarianTokenizer.from_pretrained(model_name)
127
+ model = MarianMTModel.from_pretrained(model_name)
128
+ translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
129
+
130
+ for t in translated:
131
+ print( tokenizer.decode(t, skip_special_tokens=True) )
132
+
133
+ # expected output:
134
+ # Це моє місце.
135
+ # У Тома проблемы со здоровьем.
136
+ ```
137
+
138
+ You can also use OPUS-MT models with the transformers pipelines, for example:
139
+
140
+ ```python
141
+ from transformers import pipeline
142
+ pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-pt-zle")
143
+ print(pipe(">>ukr<< Esse é o meu lugar."))
144
+
145
+ # expected output: Це моє місце.
146
+ ```
147
+
148
+ ## Benchmarks
149
+
150
+ * test set translations: [opusTCv20210807_transformer-big_2022-03-23.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/por-zle/opusTCv20210807_transformer-big_2022-03-23.test.txt)
151
+ * test set scores: [opusTCv20210807_transformer-big_2022-03-23.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/por-zle/opusTCv20210807_transformer-big_2022-03-23.eval.txt)
152
+ * benchmark results: [benchmark_results.txt](benchmark_results.txt)
153
+ * benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
154
+
155
+ | langpair | testset | chr-F | BLEU | #sent | #words |
156
+ |----------|---------|-------|-------|-------|--------|
157
+ | por-rus | tatoeba-test-v2021-08-07 | 0.67980 | 47.6 | 10000 | 65326 |
158
+ | por-ukr | tatoeba-test-v2021-08-07 | 0.65867 | 44.7 | 3372 | 18933 |
159
+ | por-rus | flores101-devtest | 0.54675 | 26.8 | 1012 | 23295 |
160
+ | por-ukr | flores101-devtest | 0.53690 | 25.1 | 1012 | 22810 |
161
+
162
+ ## Acknowledgements
163
+
164
+ The work is supported by the [European Language Grid](https://www.european-language-grid.eu/) as [pilot project 2866](https://live.european-language-grid.eu/catalogue/#/resource/projects/2866), by the [FoTran project](https://www.helsinki.fi/en/researchgroups/natural-language-understanding-with-cross-lingual-grounding), funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 771113), and the [MeMAD project](https://memad.eu/), funded by the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement No 780069. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland.
165
+
166
+ ## Model conversion info
167
+
168
+ * transformers version: 4.16.2
169
+ * OPUS-MT git hash: 1bdabf7
170
+ * port time: Thu Mar 24 03:20:20 EET 2022
171
+ * port machine: LM0-400-22516.local
benchmark_results.txt ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ por-bel flores101-dev 0.38682 8.8 997 23996
2
+ por-rus flores101-dev 0.54531 26.8 997 22657
3
+ por-ukr flores101-dev 0.52582 23.9 997 21841
4
+ por-bel flores101-devtest 0.38975 9.3 1012 24829
5
+ por-rus flores101-devtest 0.54675 26.8 1012 23295
6
+ por-ukr flores101-devtest 0.53690 25.1 1012 22810
7
+ por-rus tatoeba-test-v2021-08-07 0.67980 47.6 10000 65326
8
+ por-ukr tatoeba-test-v2021-08-07 0.65867 44.7 3372 18933
benchmark_translations.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd2bb695f858642654cc04f1a89d8fa8fe063e371d6867cb2353fd218e041cf7
3
+ size 1735345
config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "relu",
4
+ "architectures": [
5
+ "MarianMTModel"
6
+ ],
7
+ "attention_dropout": 0.0,
8
+ "bad_words_ids": [
9
+ [
10
+ 61344
11
+ ]
12
+ ],
13
+ "bos_token_id": 0,
14
+ "classifier_dropout": 0.0,
15
+ "d_model": 1024,
16
+ "decoder_attention_heads": 16,
17
+ "decoder_ffn_dim": 4096,
18
+ "decoder_layerdrop": 0.0,
19
+ "decoder_layers": 6,
20
+ "decoder_start_token_id": 61344,
21
+ "decoder_vocab_size": 61345,
22
+ "dropout": 0.1,
23
+ "encoder_attention_heads": 16,
24
+ "encoder_ffn_dim": 4096,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 6,
27
+ "eos_token_id": 27106,
28
+ "forced_eos_token_id": 27106,
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "max_length": 512,
32
+ "max_position_embeddings": 1024,
33
+ "model_type": "marian",
34
+ "normalize_embedding": false,
35
+ "num_beams": 4,
36
+ "num_hidden_layers": 6,
37
+ "pad_token_id": 61344,
38
+ "scale_embedding": true,
39
+ "share_encoder_decoder_embeddings": true,
40
+ "static_position_embeddings": true,
41
+ "torch_dtype": "float16",
42
+ "transformers_version": "4.18.0.dev0",
43
+ "use_cache": true,
44
+ "vocab_size": 61345
45
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:29de90329f122eb028f2ae451f8dc4e5c806cff0a910af9c9b4d38f2dd14dcf6
3
+ size 604198275
source.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e9a2dea0771e44c2e2d0c09ed2a6f7f1c305fef9556a4eed28cefb2921a1be69
3
+ size 823761
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}
target.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:56f30f06722151f0e1c0d99891d9a0d6946c0f4b305e87377054119bbe0735eb
3
+ size 1007525
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": "pt", "target_lang": "zle", "unk_token": "<unk>", "eos_token": "</s>", "pad_token": "<pad>", "model_max_length": 512, "sp_model_kwargs": {}, "separate_vocabs": false, "special_tokens_map_file": null, "name_or_path": "marian-models/opusTCv20210807_transformer-big_2022-03-23/pt-zle", "tokenizer_class": "MarianTokenizer"}
vocab.json ADDED
The diff for this file is too large to render. See raw diff