Initial commit

.gitattributes

@@ -29,3 +29,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.spm filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,286 @@
+---
+language:
+- da
+- de
+- gmq
+- is
+- nb
+- nn
+- no
+- sv
+
+tags:
+- translation
+- opus-mt-tc
+
+license: cc-by-4.0
+model-index:
+- name: opus-mt-tc-big-de-gmq
+  results:
+  - task:
+      name: Translation deu-dan
+      type: translation
+      args: deu-dan
+    dataset:
+      name: flores101-devtest
+      type: flores_101
+      args: deu dan devtest
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 35.6
+       - name: chr-F
+         type: chrf
+         value: 0.62363
+  - task:
+      name: Translation deu-isl
+      type: translation
+      args: deu-isl
+    dataset:
+      name: flores101-devtest
+      type: flores_101
+      args: deu isl devtest
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 20.6
+       - name: chr-F
+         type: chrf
+         value: 0.48691
+  - task:
+      name: Translation deu-nob
+      type: translation
+      args: deu-nob
+    dataset:
+      name: flores101-devtest
+      type: flores_101
+      args: deu nob devtest
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 25.3
+       - name: chr-F
+         type: chrf
+         value: 0.55765
+  - task:
+      name: Translation deu-swe
+      type: translation
+      args: deu-swe
+    dataset:
+      name: flores101-devtest
+      type: flores_101
+      args: deu swe devtest
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 34.7
+       - name: chr-F
+         type: chrf
+         value: 0.62323
+  - task:
+      name: Translation deu-dan
+      type: translation
+      args: deu-dan
+    dataset:
+      name: tatoeba-test-v2021-08-07
+      type: tatoeba_mt
+      args: deu-dan
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 58.7
+       - name: chr-F
+         type: chrf
+         value: 0.74306
+  - task:
+      name: Translation deu-isl
+      type: translation
+      args: deu-isl
+    dataset:
+      name: tatoeba-test-v2021-08-07
+      type: tatoeba_mt
+      args: deu-isl
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 47.1
+       - name: chr-F
+         type: chrf
+         value: 0.65180
+  - task:
+      name: Translation deu-nob
+      type: translation
+      args: deu-nob
+    dataset:
+      name: tatoeba-test-v2021-08-07
+      type: tatoeba_mt
+      args: deu-nob
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 52.5
+       - name: chr-F
+         type: chrf
+         value: 0.71062
+  - task:
+      name: Translation deu-swe
+      type: translation
+      args: deu-swe
+    dataset:
+      name: tatoeba-test-v2021-08-07
+      type: tatoeba_mt
+      args: deu-swe
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 58.3
+       - name: chr-F
+         type: chrf
+         value: 0.72658
+---
+# opus-mt-tc-big-de-gmq
+
+## Table of Contents
+- [Model Details](#model-details)
+- [Uses](#uses)
+- [Risks, Limitations and Biases](#risks-limitations-and-biases)
+- [How to Get Started With the Model](#how-to-get-started-with-the-model)
+- [Training](#training)
+- [Evaluation](#evaluation)
+- [Citation Information](#citation-information)
+- [Acknowledgements](#acknowledgements)
+
+## Model Details
+
+Neural machine translation model for translating from German (de) to North Germanic languages (gmq).
+
+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).
+**Model Description:**
+- **Developed by:** Language Technology Research Group at the University of Helsinki
+- **Model Type:** Translation (transformer-big)
+- **Release**: 2022-07-25
+- **License:** CC-BY-4.0
+- **Language(s):**  
+  - Source Language(s): deu
+  - Target Language(s): dan isl nno nob nor swe
+  - Valid Target Language Labels: >>dan<< >>isl<< >>nno<< >>nob<< >>nor<< >>swe<<
+- **Original Model**: [opusTCv20210807_transformer-big_2022-07-25.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/deu-gmq/opusTCv20210807_transformer-big_2022-07-25.zip)
+- **Resources for more information:**
+  - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
+  - More information about released models for this language pair: [OPUS-MT deu-gmq README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/deu-gmq/README.md)
+  - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian)
+  - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/
+
+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. `>>dan<<`
+
+## Uses
+
+This model can be used for translation and text-to-text generation.
+
+## Risks, Limitations and Biases
+
+**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.**
+
+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)).
+
+## How to Get Started With the Model
+
+A short example code:
+
+```python
+from transformers import MarianMTModel, MarianTokenizer
+
+src_text = [
+    ">>dan<< Ich hätte fast meinen Pass vergessen.",
+    ">>dan<< Dieses Fenster hier ist schusssicher."
+]
+
+model_name = "pytorch-models/opus-mt-tc-big-de-gmq"
+tokenizer = MarianTokenizer.from_pretrained(model_name)
+model = MarianMTModel.from_pretrained(model_name)
+translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
+
+for t in translated:
+    print( tokenizer.decode(t, skip_special_tokens=True) )
+
+# expected output:
+#     Jeg havde næsten glemt mit pas.
+#     Dette vindue er skudsikkert.
+```
+
+You can also use OPUS-MT models with the transformers pipelines, for example:
+
+```python
+from transformers import pipeline
+pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-de-gmq")
+print(pipe(">>dan<< Ich hätte fast meinen Pass vergessen."))
+
+# expected output: Jeg havde næsten glemt mit pas.
+```
+
+## Training
+
+- **Data**: opusTCv20210807 ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
+- **Pre-processing**: SentencePiece (spm32k,spm32k)
+- **Model Type:**  transformer-big
+- **Original MarianNMT Model**: [opusTCv20210807_transformer-big_2022-07-25.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/deu-gmq/opusTCv20210807_transformer-big_2022-07-25.zip)
+- **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
+
+## Evaluation
+
+* test set translations: [opusTCv20210807_transformer-big_2022-07-25.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/deu-gmq/opusTCv20210807_transformer-big_2022-07-25.test.txt)
+* test set scores: [opusTCv20210807_transformer-big_2022-07-25.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/deu-gmq/opusTCv20210807_transformer-big_2022-07-25.eval.txt)
+* benchmark results: [benchmark_results.txt](benchmark_results.txt)
+* benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
+
+| langpair | testset | chr-F | BLEU  | #sent | #words |
+|----------|---------|-------|-------|-------|--------|
+| deu-dan | tatoeba-test-v2021-08-07 | 0.74306 | 58.7 | 9998 | 74644 |
+| deu-isl | tatoeba-test-v2021-08-07 | 0.65180 | 47.1 | 969 | 5951 |
+| deu-nob | tatoeba-test-v2021-08-07 | 0.71062 | 52.5 | 3525 | 31978 |
+| deu-swe | tatoeba-test-v2021-08-07 | 0.72658 | 58.3 | 3410 | 22701 |
+| deu-dan | flores101-devtest | 0.62363 | 35.6 | 1012 | 24638 |
+| deu-isl | flores101-devtest | 0.48691 | 20.6 | 1012 | 22834 |
+| deu-nob | flores101-devtest | 0.55765 | 25.3 | 1012 | 23873 |
+| deu-swe | flores101-devtest | 0.62323 | 34.7 | 1012 | 23121 |
+
+## Citation Information
+
+* 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.)
+
+```
+@inproceedings{tiedemann-thottingal-2020-opus,
+    title = "{OPUS}-{MT} {--} Building open translation services for the World",
+    author = {Tiedemann, J{\"o}rg  and Thottingal, Santhosh},
+    booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
+    month = nov,
+    year = "2020",
+    address = "Lisboa, Portugal",
+    publisher = "European Association for Machine Translation",
+    url = "https://aclanthology.org/2020.eamt-1.61",
+    pages = "479--480",
+}
+
+@inproceedings{tiedemann-2020-tatoeba,
+    title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
+    author = {Tiedemann, J{\"o}rg},
+    booktitle = "Proceedings of the Fifth Conference on Machine Translation",
+    month = nov,
+    year = "2020",
+    address = "Online",
+    publisher = "Association for Computational Linguistics",
+    url = "https://aclanthology.org/2020.wmt-1.139",
+    pages = "1174--1182",
+}
+```
+
+## Acknowledgements
+
+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.
+
+## Model conversion info
+
+* transformers version: 4.16.2
+* OPUS-MT git hash: 8b9f0b0
+* port time: Fri Aug 12 18:54:58 EEST 2022
+* port machine: LM0-400-22516.local

benchmark_results.txt

@@ -0,0 +1,16 @@
+deu-dan	flores101-dev	0.61917	36.1	997	23685
+deu-isl	flores101-dev	0.48506	20.1	997	21857
+deu-nob	flores101-dev	0.55184	25.5	997	23157
+deu-swe	flores101-dev	0.62102	34.8	997	22417
+deu-dan	flores101-devtest	0.62363	35.6	1012	24638
+deu-isl	flores101-devtest	0.48691	20.6	1012	22834
+deu-nob	flores101-devtest	0.55765	25.3	1012	23873
+deu-swe	flores101-devtest	0.62323	34.7	1012	23121
+deu-dan	tatoeba-test-v2020-07-28	0.74305	58.7	10000	74655
+deu-isl	tatoeba-test-v2020-07-28	0.65276	47.2	968	5939
+deu-dan	tatoeba-test-v2021-03-30	0.74305	58.7	10000	74655
+deu-isl	tatoeba-test-v2021-03-30	0.65393	47.3	990	6127
+deu-dan	tatoeba-test-v2021-08-07	0.74306	58.7	9998	74644
+deu-isl	tatoeba-test-v2021-08-07	0.65180	47.1	969	5951
+deu-nob	tatoeba-test-v2021-08-07	0.71062	52.5	3525	31978
+deu-swe	tatoeba-test-v2021-08-07	0.72658	58.3	3410	22701

benchmark_translations.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8cd1261fb584ea7713d5dea17ddaf3fb5be08dd0f7762e278871d73da923c81e
+size 2720078

config.json

@@ -0,0 +1,45 @@
+{
+  "activation_dropout": 0.0,
+  "activation_function": "relu",
+  "architectures": [
+    "MarianMTModel"
+  ],
+  "attention_dropout": 0.0,
+  "bad_words_ids": [
+    [
+      56727
+    ]
+  ],
+  "bos_token_id": 0,
+  "classifier_dropout": 0.0,
+  "d_model": 1024,
+  "decoder_attention_heads": 16,
+  "decoder_ffn_dim": 4096,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 6,
+  "decoder_start_token_id": 56727,
+  "decoder_vocab_size": 56728,
+  "dropout": 0.1,
+  "encoder_attention_heads": 16,
+  "encoder_ffn_dim": 4096,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 6,
+  "eos_token_id": 41373,
+  "forced_eos_token_id": 41373,
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "max_length": 512,
+  "max_position_embeddings": 1024,
+  "model_type": "marian",
+  "normalize_embedding": false,
+  "num_beams": 4,
+  "num_hidden_layers": 6,
+  "pad_token_id": 56727,
+  "scale_embedding": true,
+  "share_encoder_decoder_embeddings": true,
+  "static_position_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.18.0.dev0",
+  "use_cache": true,
+  "vocab_size": 56728
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db3fcc4f08b00f84532ba66c66511b3be3649a81f34af2516079e8ea7e0dcbe4
+size 585277763

source.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06a1b553146ea2267fa9c35c807334729e5da52891fed3f4586543db9dd18c37
+size 823934

special_tokens_map.json

@@ -0,0 +1 @@
+{"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}

target.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abe1d5af0e56a45bcd1310b56ea9d5c571caab24e2d319556b3f10e414824fdd
+size 805921

tokenizer_config.json

@@ -0,0 +1 @@
+{"source_lang": "de", "target_lang": "gmq", "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_2022-07-25/de-gmq", "tokenizer_class": "MarianTokenizer"}