tiedeman commited on
Commit
3f71b6b
1 Parent(s): fa8e8a2

Initial commit

Browse files
README.md ADDED
@@ -0,0 +1,67 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ - tr
5
+
6
+ tags:
7
+ - translation
8
+
9
+ license: apache-2.0
10
+ ---
11
+ ### en-tr
12
+
13
+ * source group: English
14
+ * target group: Turkish
15
+ * OPUS readme: [eng-tur](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/eng-tur/README.md)
16
+
17
+ * model: transformer-align
18
+ * source language(s): eng
19
+ * target language(s): tur
20
+ * model: transformer-align
21
+ * pre-processing: normalization + SentencePiece (spm32k,spm32k)
22
+ * download original weights: [opus+bt-2021-04-10.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/eng-tur/opus+bt-2021-04-10.zip)
23
+ * test set translations: [opus+bt-2021-04-10.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/eng-tur/opus+bt-2021-04-10.test.txt)
24
+ * test set scores: [opus+bt-2021-04-10.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/eng-tur/opus+bt-2021-04-10.eval.txt)
25
+
26
+ ## Benchmarks
27
+
28
+ | testset | BLEU | chr-F | #sent | #words | BP |
29
+ |---------|-------|-------|-------|--------|----|
30
+ | newsdev2016-entr.eng-tur | 21.5 | 0.575 | 1001 | 16127 | 1.000 |
31
+ | newstest2016-entr.eng-tur | 21.4 | 0.558 | 3000 | 50782 | 0.986 |
32
+ | newstest2017-entr.eng-tur | 22.8 | 0.572 | 3007 | 51977 | 0.960 |
33
+ | newstest2018-entr.eng-tur | 20.8 | 0.561 | 3000 | 53731 | 0.963 |
34
+ | Tatoeba-test.eng-tur | 41.5 | 0.684 | 10000 | 60469 | 0.932 |
35
+
36
+
37
+ ### System Info:
38
+ - hf_name: en-tr
39
+ - source_languages: eng
40
+ - target_languages: tur
41
+ - opus_readme_url: https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/eng-tur/README.md
42
+ - original_repo: Tatoeba-Challenge
43
+ - tags: ['translation']
44
+ - languages: ['en', 'tr']
45
+ - src_constituents: ('English', {'eng'})
46
+ - tgt_constituents: ('Turkish', {'tur'})
47
+ - src_multilingual: False
48
+ - tgt_multilingual: False
49
+ - long_pair: eng-tur
50
+ - prepro: normalization + SentencePiece (spm32k,spm32k)
51
+ - url_model: https://object.pouta.csc.fi/Tatoeba-MT-models/eng-tur/opus+bt-2021-04-10.zip
52
+ - url_test_set: https://object.pouta.csc.fi/Tatoeba-MT-models/eng-tur/opus+bt-2021-04-10.test.txt
53
+ - src_alpha3: eng
54
+ - tgt_alpha3: tur
55
+ - chrF2_score: 0.684
56
+ - bleu: 41.5
57
+ - src_name: English
58
+ - tgt_name: Turkish
59
+ - train_date: 2021-04-10 00:00:00
60
+ - src_alpha2: en
61
+ - tgt_alpha2: tr
62
+ - prefer_old: False
63
+ - short_pair: en-tr
64
+ - helsinki_git_sha: a6bd0607aec9603811b2b635aec3f566f3add79d
65
+ - transformers_git_sha: 12b4d66a80419db30a15e7b9d4208ceb9887c03b
66
+ - port_machine: LM0-400-22516.local
67
+ - port_time: 2021-10-05-12:13
config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "swish",
4
+ "architectures": [
5
+ "MarianMTModel"
6
+ ],
7
+ "attention_dropout": 0.0,
8
+ "bad_words_ids": [
9
+ [
10
+ 59993
11
+ ]
12
+ ],
13
+ "bos_token_id": 0,
14
+ "classifier_dropout": 0.0,
15
+ "d_model": 512,
16
+ "decoder_attention_heads": 8,
17
+ "decoder_ffn_dim": 2048,
18
+ "decoder_layerdrop": 0.0,
19
+ "decoder_layers": 6,
20
+ "decoder_start_token_id": 59993,
21
+ "dropout": 0.1,
22
+ "encoder_attention_heads": 8,
23
+ "encoder_ffn_dim": 2048,
24
+ "encoder_layerdrop": 0.0,
25
+ "encoder_layers": 6,
26
+ "eos_token_id": 0,
27
+ "forced_eos_token_id": 0,
28
+ "init_std": 0.02,
29
+ "is_encoder_decoder": true,
30
+ "max_length": 512,
31
+ "max_position_embeddings": 512,
32
+ "model_type": "marian",
33
+ "normalize_embedding": false,
34
+ "num_beams": 6,
35
+ "num_hidden_layers": 6,
36
+ "pad_token_id": 59993,
37
+ "scale_embedding": true,
38
+ "static_position_embeddings": true,
39
+ "torch_dtype": "float16",
40
+ "transformers_version": "4.12.0.dev0",
41
+ "use_cache": true,
42
+ "vocab_size": 59994
43
+ }
metadata.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"hf_name":"en-tr","source_languages":"eng","target_languages":"tur","opus_readme_url":"https:\/\/github.com\/Helsinki-NLP\/Tatoeba-Challenge\/tree\/master\/models\/eng-tur\/README.md","original_repo":"Tatoeba-Challenge","tags":["translation"],"languages":["en","tr"],"src_constituents":["English",["eng"]],"tgt_constituents":["Turkish",["tur"]],"src_multilingual":false,"tgt_multilingual":false,"long_pair":"eng-tur","prepro":" normalization + SentencePiece (spm32k,spm32k)","url_model":"https:\/\/object.pouta.csc.fi\/Tatoeba-MT-models\/eng-tur\/opus+bt-2021-04-10.zip","url_test_set":"https:\/\/object.pouta.csc.fi\/Tatoeba-MT-models\/eng-tur\/opus+bt-2021-04-10.test.txt","src_alpha3":"eng","tgt_alpha3":"tur","chrF2_score":0.684,"bleu":41.5,"src_name":"English","tgt_name":"Turkish","train_date":1618012800000,"src_alpha2":"en","tgt_alpha2":"tr","prefer_old":false,"short_pair":"en-tr","helsinki_git_sha":"a6bd0607aec9603811b2b635aec3f566f3add79d","transformers_git_sha":"12b4d66a80419db30a15e7b9d4208ceb9887c03b","port_machine":"LM0-400-22516.local","port_time":"2021-10-05-12:13"}
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07454925f30337579bb8e3edc3662f6a697386cd9dadd9fba4eddfe4296d2459
3
+ size 211352677
source.spm ADDED
Binary file (794 kB). View file
 
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}
target.spm ADDED
Binary file (838 kB). View file
 
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": "eng", "target_lang": "tur", "unk_token": "<unk>", "eos_token": "</s>", "pad_token": "<pad>", "model_max_length": 512, "sp_model_kwargs": {}, "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "marian_ckpt/eng-tur", "tokenizer_class": "MarianTokenizer"}
vocab.json ADDED
The diff for this file is too large to render. See raw diff