Initial commit

.gitattributes

@@ -26,3 +26,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zstandard 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 +1,323 @@
+---
+language:
+- ces+slk
+- cs
+- en
+
+tags:
+- translation
+
+license: cc-by-4.0
+model-index:
+- name: opus-mt-tc-big-ces_slk-en
+  results:
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: flores101-devtest
+      type: flores_101
+      args: ces eng devtest
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 41.2
+  - task:
+      name: Translation slk-eng
+      type: translation
+      args: slk-eng
+    dataset:
+      name: flores101-devtest
+      type: flores_101
+      args: slk eng devtest
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 40.1
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: multi30k_test_2016_flickr
+      type: multi30k-2016_flickr
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 38.6
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: multi30k_test_2018_flickr
+      type: multi30k-2018_flickr
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 37.9
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: news-test2008
+      type: news-test2008
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 26.2
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: tatoeba-test-v2021-08-07
+      type: tatoeba_mt
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 57.7
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2009
+      type: wmt-2009-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 28.8
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2010
+      type: wmt-2010-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 30.3
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2011
+      type: wmt-2011-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 30.3
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2012
+      type: wmt-2012-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 29.4
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2013
+      type: wmt-2013-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 33.1
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2014
+      type: wmt-2014-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 38.3
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2015
+      type: wmt-2015-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 33.6
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2016
+      type: wmt-2016-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 36.8
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2017
+      type: wmt-2017-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 32.3
+  - task:
+      name: Translation ces-eng
+      type: translation
+      args: ces-eng
+    dataset:
+      name: newstest2018
+      type: wmt-2018-news
+      args: ces-eng
+    metrics:
+       - name: BLEU
+         type: bleu
+         value: 33.0
+---
+# opus-mt-tc-big-ces_slk-en
+
+Neural machine translation model for translating from unknown (ces+slk) to English (en).
+
+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).
+
+* 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",
+}
+```
+
+## Model info
+
+* Release: 2022-03-17
+* source language(s): ces
+* target language(s): eng
+* model: transformer-big
+* data: opusTCv20210807+bt ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
+* tokenization: SentencePiece (spm32k,spm32k)
+* original model: [opusTCv20210807+bt_transformer-big_2022-03-17.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/ces+slk-eng/opusTCv20210807+bt_transformer-big_2022-03-17.zip)
+* more information released models: [OPUS-MT ces+slk-eng README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/ces+slk-eng/README.md)
+
+## Usage
+
+A short example code:
+
+```python
+from transformers import MarianMTModel, MarianTokenizer
+
+src_text = [
+    "Podívej se na své kalhoty! Zapni si je na zip.",
+    "Mrzí mě, že Tom odchází."
+]
+
+model_name = "pytorch-models/opus-mt-tc-big-ces_slk-en"
+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:
+#     Look at your pants, zip them up.
+#     I'm sorry Tom's leaving.
+```
+
+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-ces_slk-en")
+print(pipe("Podívej se na své kalhoty! Zapni si je na zip."))
+
+# expected output: Look at your pants, zip them up.
+```
+
+## Benchmarks
+
+* test set translations: [opusTCv20210807+bt_transformer-big_2022-03-17.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/ces+slk-eng/opusTCv20210807+bt_transformer-big_2022-03-17.test.txt)
+* test set scores: [opusTCv20210807+bt_transformer-big_2022-03-17.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/ces+slk-eng/opusTCv20210807+bt_transformer-big_2022-03-17.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 |
+|----------|---------|-------|-------|-------|--------|
+| ces-eng | tatoeba-test-v2021-08-07 | 0.72120 | 57.7 | 13824 | 105010 |
+| ces-eng | flores101-devtest | 0.66511 | 41.2 | 1012 | 24721 |
+| slk-eng | flores101-devtest | 0.66084 | 40.1 | 1012 | 24721 |
+| ces-eng | multi30k_test_2016_flickr | 0.62216 | 38.6 | 1000 | 12955 |
+| ces-eng | multi30k_test_2018_flickr | 0.61838 | 37.9 | 1071 | 14689 |
+| ces-eng | newssyscomb2009 | 0.56380 | 29.9 | 502 | 11818 |
+| ces-eng | news-test2008 | 0.54071 | 26.2 | 2051 | 49380 |
+| ces-eng | newstest2009 | 0.55871 | 28.8 | 2525 | 65399 |
+| ces-eng | newstest2010 | 0.57634 | 30.3 | 2489 | 61711 |
+| ces-eng | newstest2011 | 0.57002 | 30.3 | 3003 | 74681 |
+| ces-eng | newstest2012 | 0.56564 | 29.4 | 3003 | 72812 |
+| ces-eng | newstest2013 | 0.58723 | 33.1 | 3000 | 64505 |
+| ces-eng | newstest2014 | 0.64192 | 38.3 | 3003 | 68065 |
+| ces-eng | newstest2015 | 0.58688 | 33.6 | 2656 | 53569 |
+| ces-eng | newstest2016 | 0.61544 | 36.8 | 2999 | 64670 |
+| ces-eng | newstest2017 | 0.58085 | 32.3 | 3005 | 61721 |
+| ces-eng | newstest2018 | 0.58627 | 33.0 | 2983 | 63495 |
+
+## 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: 3405783
+* port time: Wed Apr 13 18:42:25 EEST 2022
+* port machine: LM0-400-22516.local

benchmark_results.txt

@@ -0,0 +1,21 @@
+ces-eng	flores101-dev	0.65682	40.3	997	23555
+slk-eng	flores101-dev	0.66564	40.5	997	23555
+ces-eng	flores101-devtest	0.66511	41.2	1012	24721
+slk-eng	flores101-devtest	0.66084	40.1	1012	24721
+ces-eng	multi30k_test_2016_flickr	0.62216	38.6	1000	12955
+ces-eng	multi30k_test_2018_flickr	0.61838	37.9	1071	14689
+ces-eng	newssyscomb2009	0.56380	29.9	502	11818
+ces-eng	news-test2008	0.54071	26.2	2051	49380
+ces-eng	newstest2009	0.55871	28.8	2525	65399
+ces-eng	newstest2010	0.57634	30.3	2489	61711
+ces-eng	newstest2011	0.57002	30.3	3003	74681
+ces-eng	newstest2012	0.56564	29.4	3003	72812
+ces-eng	newstest2013	0.58723	33.1	3000	64505
+ces-eng	newstest2014	0.64192	38.3	3003	68065
+ces-eng	newstest2015	0.58688	33.6	2656	53569
+ces-eng	newstest2016	0.61544	36.8	2999	64670
+ces-eng	newstest2017	0.58085	32.3	3005	61721
+ces-eng	newstest2018	0.58627	33.0	2983	63495
+ces-eng	tatoeba-test-v2020-07-28	0.72914	58.7	10000	75376
+ces-eng	tatoeba-test-v2021-03-30	0.72430	58.1	12076	91333
+ces-eng	tatoeba-test-v2021-08-07	0.72120	57.7	13824	105010

benchmark_translations.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:219badbde1aff531be2d8fec8be7b0423e7de76b38c0ae0022fc11dbf9d53561
+size 6159479

config.json

@@ -0,0 +1,45 @@
+{
+  "activation_dropout": 0.0,
+  "activation_function": "relu",
+  "architectures": [
+    "MarianMTModel"
+  ],
+  "attention_dropout": 0.0,
+  "bad_words_ids": [
+    [
+      57713
+    ]
+  ],
+  "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": 57713,
+  "decoder_vocab_size": 57714,
+  "dropout": 0.1,
+  "encoder_attention_heads": 16,
+  "encoder_ffn_dim": 4096,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 6,
+  "eos_token_id": 42506,
+  "forced_eos_token_id": 42506,
+  "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": 57713,
+  "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": 57714
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8b2b661c4df5163032e90aeb3e431efe1019615058270a6a251625a60a1ff46
+size 589318403

source.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa297556cf75acf29f51dea0f1c0ef5225d247450e05307fd44dea6ea3185251
+size 823611

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:b1c603a8ddb42de2558898ca8d1ff49eb9549bc15d5922102c9890a2db4108f5
+size 796116

tokenizer_config.json

@@ -0,0 +1 @@
+{"source_lang": "ces+slk", "target_lang": "en", "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_2022-03-17/ces+slk-en", "tokenizer_class": "MarianTokenizer"}