tiedeman commited on
Commit
8510db8
1 Parent(s): 1a77b82

Initial commit

Browse files
.gitattributes CHANGED
@@ -26,3 +26,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
26
  *.zip filter=lfs diff=lfs merge=lfs -text
27
  *.zstandard filter=lfs diff=lfs merge=lfs -text
28
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
26
  *.zip filter=lfs diff=lfs merge=lfs -text
27
  *.zstandard filter=lfs diff=lfs merge=lfs -text
28
  *tfevents* filter=lfs diff=lfs merge=lfs -text
29
+ *.spm filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,144 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ - pt
5
+ - pt_br
6
+
7
+ tags:
8
+ - translation
9
+
10
+ license: cc-by-4.0
11
+ model-index:
12
+ - name: opus-mt-tc-big-en-pt
13
+ results:
14
+ - task:
15
+ name: Translation eng-por
16
+ type: translation
17
+ args: eng-por
18
+ dataset:
19
+ name: flores101-devtest
20
+ type: flores_101
21
+ args: eng por devtest
22
+ metrics:
23
+ - name: BLEU
24
+ type: bleu
25
+ value: 50.4
26
+ - task:
27
+ name: Translation eng-por
28
+ type: translation
29
+ args: eng-por
30
+ dataset:
31
+ name: tatoeba-test-v2021-08-07
32
+ type: tatoeba_mt
33
+ args: eng-por
34
+ metrics:
35
+ - name: BLEU
36
+ type: bleu
37
+ value: 49.6
38
+ ---
39
+ # opus-mt-tc-big-en-pt
40
+
41
+ Neural machine translation model for translating from English (en) to Portuguese (pt).
42
+
43
+ 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).
44
+
45
+ * 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.)
46
+
47
+ ```
48
+ @inproceedings{tiedemann-thottingal-2020-opus,
49
+ title = "{OPUS}-{MT} {--} Building open translation services for the World",
50
+ author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh},
51
+ booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
52
+ month = nov,
53
+ year = "2020",
54
+ address = "Lisboa, Portugal",
55
+ publisher = "European Association for Machine Translation",
56
+ url = "https://aclanthology.org/2020.eamt-1.61",
57
+ pages = "479--480",
58
+ }
59
+
60
+ @inproceedings{tiedemann-2020-tatoeba,
61
+ title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
62
+ author = {Tiedemann, J{\"o}rg},
63
+ booktitle = "Proceedings of the Fifth Conference on Machine Translation",
64
+ month = nov,
65
+ year = "2020",
66
+ address = "Online",
67
+ publisher = "Association for Computational Linguistics",
68
+ url = "https://aclanthology.org/2020.wmt-1.139",
69
+ pages = "1174--1182",
70
+ }
71
+ ```
72
+
73
+ ## Model info
74
+
75
+ * Release: 2022-03-13
76
+ * source language(s): eng
77
+ * target language(s): pob por
78
+ * valid target language labels: >>pob<< >>por<<
79
+ * model: transformer-big
80
+ * data: opusTCv20210807+bt ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
81
+ * tokenization: SentencePiece (spm32k,spm32k)
82
+ * original model: [opusTCv20210807+bt_transformer-big_2022-03-13.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/eng-por/opusTCv20210807+bt_transformer-big_2022-03-13.zip)
83
+ * more information released models: [OPUS-MT eng-por README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/eng-por/README.md)
84
+ * more information about the model: [MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)
85
+
86
+ 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. `>>pob<<`
87
+
88
+ ## Usage
89
+
90
+ A short example code:
91
+
92
+ ```python
93
+ from transformers import MarianMTModel, MarianTokenizer
94
+
95
+ src_text = [
96
+ ">>por<< Tom tried to stab me.",
97
+ ">>por<< He has been to Hawaii several times."
98
+ ]
99
+
100
+ model_name = "pytorch-models/opus-mt-tc-big-en-pt"
101
+ tokenizer = MarianTokenizer.from_pretrained(model_name)
102
+ model = MarianMTModel.from_pretrained(model_name)
103
+ translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
104
+
105
+ for t in translated:
106
+ print( tokenizer.decode(t, skip_special_tokens=True) )
107
+
108
+ # expected output:
109
+ # O Tom tentou esfaquear-me.
110
+ # Ele já esteve no Havaí várias vezes.
111
+ ```
112
+
113
+ You can also use OPUS-MT models with the transformers pipelines, for example:
114
+
115
+ ```python
116
+ from transformers import pipeline
117
+ pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-en-pt")
118
+ print(pipe(">>por<< Tom tried to stab me."))
119
+
120
+ # expected output: O Tom tentou esfaquear-me.
121
+ ```
122
+
123
+ ## Benchmarks
124
+
125
+ * test set translations: [opusTCv20210807+bt_transformer-big_2022-03-13.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/eng-por/opusTCv20210807+bt_transformer-big_2022-03-13.test.txt)
126
+ * test set scores: [opusTCv20210807+bt_transformer-big_2022-03-13.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/eng-por/opusTCv20210807+bt_transformer-big_2022-03-13.eval.txt)
127
+ * benchmark results: [benchmark_results.txt](benchmark_results.txt)
128
+ * benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
129
+
130
+ | langpair | testset | chr-F | BLEU | #sent | #words |
131
+ |----------|---------|-------|-------|-------|--------|
132
+ | eng-por | tatoeba-test-v2021-08-07 | 0.69320 | 49.6 | 13222 | 105265 |
133
+ | eng-por | flores101-devtest | 0.71673 | 50.4 | 1012 | 26519 |
134
+
135
+ ## Acknowledgements
136
+
137
+ 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.
138
+
139
+ ## Model conversion info
140
+
141
+ * transformers version: 4.16.2
142
+ * OPUS-MT git hash: 3405783
143
+ * port time: Wed Apr 13 17:48:54 EEST 2022
144
+ * port machine: LM0-400-22516.local
benchmark_results.txt ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ eng-por flores101-dev 0.71108 50.0 997 25287
2
+ eng-por flores101-devtest 0.71673 50.4 1012 26519
3
+ eng-por tatoeba-test-v2020-07-28 0.68656 48.8 10000 75371
4
+ eng-por tatoeba-test-v2021-03-30 0.68785 48.9 11574 87572
5
+ eng-por tatoeba-test-v2021-08-07 0.69320 49.6 13222 105265
benchmark_translations.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3539b408a2285cd7e35a58128a29f810ad72632d3cb3de8c522cfaa5ac723be6
3
+ size 1523832
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
+ 54775
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": 54775,
21
+ "decoder_vocab_size": 54776,
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": 44670,
28
+ "forced_eos_token_id": 44670,
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": 54775,
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": 54776
45
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3f6182f219942bbd90742f0976edda92027f219d7f9c6d3904915e2673e0738
3
+ size 577278467
source.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a7fcf812cf03a5785daa35d4932bbbe69e7e605c0fe56fce5a3f731d6c355aa
3
+ size 802741
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:b3cf20ccd4e48a89882cfdba157c05fa1cd71a6b69517dfc80e3a34d387ea91f
3
+ size 824855
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": "en", "target_lang": "pt", "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+bt_transformer-big_2022-03-13/en-pt", "tokenizer_class": "MarianTokenizer"}
vocab.json ADDED
The diff for this file is too large to render. See raw diff