tiedeman commited on
Commit
052aa93
1 Parent(s): 4d6b7f4

Initial commit

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ *.spm filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,235 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ language:
4
+ - ace
5
+ - ban
6
+ - bcl
7
+ - btd
8
+ - bth
9
+ - bto
10
+ - bts
11
+ - btx
12
+ - bug
13
+ - ceb
14
+ - cgc
15
+ - ch
16
+ - dtp
17
+ - es
18
+ - fil
19
+ - fr
20
+ - gor
21
+ - hil
22
+ - iba
23
+ - id
24
+ - ify
25
+ - ilo
26
+ - it
27
+ - jv
28
+ - krj
29
+ - ljp
30
+ - mad
31
+ - mak
32
+ - mbb
33
+ - mbt
34
+ - mg
35
+ - mhy
36
+ - mog
37
+ - mrw
38
+ - ms
39
+ - msm
40
+ - mta
41
+ - mwv
42
+ - nia
43
+ - nij
44
+ - obo
45
+ - pag
46
+ - pam
47
+ - pau
48
+ - pt
49
+ - rej
50
+ - sas
51
+ - sda
52
+ - sml
53
+ - su
54
+ - sxn
55
+ - tbl
56
+ - war
57
+
58
+ tags:
59
+ - translation
60
+ - opus-mt-tc-bible
61
+
62
+ license: apache-2.0
63
+ model-index:
64
+ - name: opus-mt-tc-bible-big-pqw-fra_ita_por_spa
65
+ results:
66
+ - task:
67
+ name: Translation multi-multi
68
+ type: translation
69
+ args: multi-multi
70
+ dataset:
71
+ name: tatoeba-test-v2020-07-28-v2023-09-26
72
+ type: tatoeba_mt
73
+ args: multi-multi
74
+ metrics:
75
+ - name: BLEU
76
+ type: bleu
77
+ value: 35.3
78
+ - name: chr-F
79
+ type: chrf
80
+ value: 0.57349
81
+ ---
82
+ # opus-mt-tc-bible-big-pqw-fra_ita_por_spa
83
+
84
+ ## Table of Contents
85
+ - [Model Details](#model-details)
86
+ - [Uses](#uses)
87
+ - [Risks, Limitations and Biases](#risks-limitations-and-biases)
88
+ - [How to Get Started With the Model](#how-to-get-started-with-the-model)
89
+ - [Training](#training)
90
+ - [Evaluation](#evaluation)
91
+ - [Citation Information](#citation-information)
92
+ - [Acknowledgements](#acknowledgements)
93
+
94
+ ## Model Details
95
+
96
+ Neural machine translation model for translating from Western Malayo-Polynesian languages (pqw) to unknown (fra+ita+por+spa).
97
+
98
+ 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).
99
+ **Model Description:**
100
+ - **Developed by:** Language Technology Research Group at the University of Helsinki
101
+ - **Model Type:** Translation (transformer-big)
102
+ - **Release**: 2024-08-17
103
+ - **License:** Apache-2.0
104
+ - **Language(s):**
105
+ - Source Language(s): ace ban bcl btd bth bto bts btx bug ceb cgc cha dtp fil gor hil iba ify ilo ind jak jav krj ljp mad mak mbb mbt mhy mlg mog mrw msa msm mta mwv nia nij obo pag pam pau plt rej sas sda sml sun sxn tbl war zsm
106
+ - Target Language(s): fra ita por spa
107
+ - Valid Target Language Labels: >>fra<< >>ita<< >>por<< >>spa<< >>xxx<<
108
+ - **Original Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/pqw-fra+ita+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17.zip)
109
+ - **Resources for more information:**
110
+ - [OPUS-MT dashboard](https://opus.nlpl.eu/dashboard/index.php?pkg=opusmt&test=all&scoreslang=all&chart=standard&model=Tatoeba-MT-models/pqw-fra%2Bita%2Bpor%2Bspa/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-08-17)
111
+ - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
112
+ - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian)
113
+ - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/)
114
+ - [HPLT bilingual data v1 (as part of the Tatoeba Translation Challenge dataset)](https://hplt-project.org/datasets/v1)
115
+ - [A massively parallel Bible corpus](https://aclanthology.org/L14-1215/)
116
+
117
+ 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. `>>fra<<`
118
+
119
+ ## Uses
120
+
121
+ This model can be used for translation and text-to-text generation.
122
+
123
+ ## Risks, Limitations and Biases
124
+
125
+ **CONTENT WARNING: Readers should be aware that the model is trained on various public data sets that may contain content that is disturbing, offensive, and can propagate historical and current stereotypes.**
126
+
127
+ Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)).
128
+
129
+ ## How to Get Started With the Model
130
+
131
+ A short example code:
132
+
133
+ ```python
134
+ from transformers import MarianMTModel, MarianTokenizer
135
+
136
+ src_text = [
137
+ ">>fra<< Bag-ong iroy akong gusto.",
138
+ ">>fra<< Usá, duhá, tuló, upat, limá, unom, pitó, waló, siyam, napúlò."
139
+ ]
140
+
141
+ model_name = "pytorch-models/opus-mt-tc-bible-big-pqw-fra_ita_por_spa"
142
+ tokenizer = MarianTokenizer.from_pretrained(model_name)
143
+ model = MarianMTModel.from_pretrained(model_name)
144
+ translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
145
+
146
+ for t in translated:
147
+ print( tokenizer.decode(t, skip_special_tokens=True) )
148
+
149
+ # expected output:
150
+ # Ma mère, ma fille, c'est ma mère.
151
+ # Le cinquième, de sardonix ; le sixième, de sarde ; le septième, de chrysolithe ; le huitième, de béryl ; le neuvième, de topaze ; le dixième, de chrysoprase ; le onzième, d'hyacinthe, et le douzième, d'améthyste ;
152
+ ```
153
+
154
+ You can also use OPUS-MT models with the transformers pipelines, for example:
155
+
156
+ ```python
157
+ from transformers import pipeline
158
+ pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-bible-big-pqw-fra_ita_por_spa")
159
+ print(pipe(">>fra<< Bag-ong iroy akong gusto."))
160
+
161
+ # expected output: Ma mère, ma fille, c'est ma mère.
162
+ ```
163
+
164
+ ## Training
165
+
166
+ - **Data**: opusTCv20230926max50+bt+jhubc ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
167
+ - **Pre-processing**: SentencePiece (spm32k,spm32k)
168
+ - **Model Type:** transformer-big
169
+ - **Original MarianNMT Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/pqw-fra+ita+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17.zip)
170
+ - **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
171
+
172
+ ## Evaluation
173
+
174
+ * [Model scores at the OPUS-MT dashboard](https://opus.nlpl.eu/dashboard/index.php?pkg=opusmt&test=all&scoreslang=all&chart=standard&model=Tatoeba-MT-models/pqw-fra%2Bita%2Bpor%2Bspa/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-08-17)
175
+ * test set translations: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/pqw-fra+ita+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17.test.txt)
176
+ * test set scores: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/pqw-fra+ita+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17.eval.txt)
177
+ * benchmark results: [benchmark_results.txt](benchmark_results.txt)
178
+ * benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
179
+
180
+ | langpair | testset | chr-F | BLEU | #sent | #words |
181
+ |----------|---------|-------|-------|-------|--------|
182
+ | multi-multi | tatoeba-test-v2020-07-28-v2023-09-26 | 0.57349 | 35.3 | 2094 | 15178 |
183
+
184
+ ## Citation Information
185
+
186
+ * Publications: [Democratizing neural machine translation with OPUS-MT](https://doi.org/10.1007/s10579-023-09704-w) and [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.)
187
+
188
+ ```bibtex
189
+ @article{tiedemann2023democratizing,
190
+ title={Democratizing neural machine translation with {OPUS-MT}},
191
+ author={Tiedemann, J{\"o}rg and Aulamo, Mikko and Bakshandaeva, Daria and Boggia, Michele and Gr{\"o}nroos, Stig-Arne and Nieminen, Tommi and Raganato, Alessandro and Scherrer, Yves and Vazquez, Raul and Virpioja, Sami},
192
+ journal={Language Resources and Evaluation},
193
+ number={58},
194
+ pages={713--755},
195
+ year={2023},
196
+ publisher={Springer Nature},
197
+ issn={1574-0218},
198
+ doi={10.1007/s10579-023-09704-w}
199
+ }
200
+
201
+ @inproceedings{tiedemann-thottingal-2020-opus,
202
+ title = "{OPUS}-{MT} {--} Building open translation services for the World",
203
+ author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh},
204
+ booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
205
+ month = nov,
206
+ year = "2020",
207
+ address = "Lisboa, Portugal",
208
+ publisher = "European Association for Machine Translation",
209
+ url = "https://aclanthology.org/2020.eamt-1.61",
210
+ pages = "479--480",
211
+ }
212
+
213
+ @inproceedings{tiedemann-2020-tatoeba,
214
+ title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
215
+ author = {Tiedemann, J{\"o}rg},
216
+ booktitle = "Proceedings of the Fifth Conference on Machine Translation",
217
+ month = nov,
218
+ year = "2020",
219
+ address = "Online",
220
+ publisher = "Association for Computational Linguistics",
221
+ url = "https://aclanthology.org/2020.wmt-1.139",
222
+ pages = "1174--1182",
223
+ }
224
+ ```
225
+
226
+ ## Acknowledgements
227
+
228
+ The work is supported by the [HPLT project](https://hplt-project.org/), funded by the European Union’s Horizon Europe research and innovation programme under grant agreement No 101070350. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland, and the [EuroHPC supercomputer LUMI](https://www.lumi-supercomputer.eu/).
229
+
230
+ ## Model conversion info
231
+
232
+ * transformers version: 4.45.1
233
+ * OPUS-MT git hash: 0882077
234
+ * port time: Tue Oct 8 14:05:22 EEST 2024
235
+ * port machine: LM0-400-22516.local
benchmark_results.txt ADDED
@@ -0,0 +1 @@
 
 
1
+ multi-multi tatoeba-test-v2020-07-28-v2023-09-26 0.57349 35.3 2094 15178
benchmark_translations.zip ADDED
File without changes
config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "pytorch-models/opus-mt-tc-bible-big-pqw-fra_ita_por_spa",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "relu",
5
+ "architectures": [
6
+ "MarianMTModel"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 0,
10
+ "classifier_dropout": 0.0,
11
+ "d_model": 1024,
12
+ "decoder_attention_heads": 16,
13
+ "decoder_ffn_dim": 4096,
14
+ "decoder_layerdrop": 0.0,
15
+ "decoder_layers": 6,
16
+ "decoder_start_token_id": 60728,
17
+ "decoder_vocab_size": 60729,
18
+ "dropout": 0.1,
19
+ "encoder_attention_heads": 16,
20
+ "encoder_ffn_dim": 4096,
21
+ "encoder_layerdrop": 0.0,
22
+ "encoder_layers": 6,
23
+ "eos_token_id": 242,
24
+ "forced_eos_token_id": null,
25
+ "init_std": 0.02,
26
+ "is_encoder_decoder": true,
27
+ "max_length": null,
28
+ "max_position_embeddings": 1024,
29
+ "model_type": "marian",
30
+ "normalize_embedding": false,
31
+ "num_beams": null,
32
+ "num_hidden_layers": 6,
33
+ "pad_token_id": 60728,
34
+ "scale_embedding": true,
35
+ "share_encoder_decoder_embeddings": true,
36
+ "static_position_embeddings": true,
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.45.1",
39
+ "use_cache": true,
40
+ "vocab_size": 60729
41
+ }
generation_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bad_words_ids": [
4
+ [
5
+ 60728
6
+ ]
7
+ ],
8
+ "bos_token_id": 0,
9
+ "decoder_start_token_id": 60728,
10
+ "eos_token_id": 242,
11
+ "forced_eos_token_id": 242,
12
+ "max_length": 512,
13
+ "num_beams": 4,
14
+ "pad_token_id": 60728,
15
+ "transformers_version": "4.45.1"
16
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7700db2b23ff620790a038aa90f8f0df058342c671a20ecf5c53614ffc5aee81
3
+ size 954448020
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eef6eec3ef7c4524e164560b71bd4f402f6be1b8f8b1905710374a61115c8b95
3
+ size 954499269
source.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:75eb8e9b06981c71a572ad5e2cdbc990c4bfa532c35ca352ab717f4803e37f51
3
+ size 789110
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:c8a7c7e5b0bbc04cd9adc2f6065023024fd013d5779aa38d24885fe785109da9
3
+ size 824731
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": "pqw", "target_lang": "fra+ita+por+spa", "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/opusTCv20230926max50+bt+jhubc_transformer-big_2024-08-17/pqw-fra+ita+por+spa", "tokenizer_class": "MarianTokenizer"}
vocab.json ADDED
The diff for this file is too large to render. See raw diff