valhalla commited on
Commit
056d0cc
1 Parent(s): 28bc70e
README.md ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language: multilingual
3
+ tags:
4
+ - mbart-50
5
+ ---
6
+
7
+ # mBART-50 many to many multilingual machine translation
8
+
9
+
10
+ This model is a fine-tuned checkpoint of [mBART-large-50](https://huggingface.co/facebook/mbart-large-50). `mbart-large-50-many-to-many-mmt` is fine-tuned for multilingual machine translation. It was introduced in [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) paper.
11
+
12
+
13
+ The model can translate directly between any pair of 50 languages. To translate into a target language, the target language id is forced as the first generated token. To force the target language id as the first generated token, pass the `forced_bos_token_id` parameter to the `generate` method.
14
+
15
+
16
+ ```python
17
+ from transformers import MBartForConditionalGeneration, MBart50TokenizerFast
18
+
19
+ article_hi = "संयुक्त राष्ट्र के प्रमुख का कहना है कि सीरिया में कोई सैन्य समाधान नहीं है"
20
+ article_ar = "الأمين العام للأمم المتحدة يقول إنه لا يوجد حل عسكري في سوريا."
21
+
22
+ model = MBartForConditionalGeneration.from_pretrained("facebook/mbart-large-50-many-to-many-mmt")
23
+ tokenizer = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50-many-to-many-mmt")
24
+
25
+ # translate Hindi to French
26
+ tokenizer.src_lang = "hi_IN"
27
+ encoded_hi = tokenizer(article_hi, return_tensors="pt")
28
+ generated_tokens = model.generate(
29
+ **encoded_hi,
30
+ forced_bos_token_id=tokenizer.lang_code_to_id["fr_XX"]
31
+ )
32
+ tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
33
+ # => "Le chef de l 'ONU affirme qu 'il n 'y a pas de solution militaire dans la Syrie."
34
+
35
+ # translate Arabic to English
36
+ tokenizer.src_lang = "ar_AR"
37
+ encoded_ar = tokenizer(article_ar, return_tensors="pt")
38
+ generated_tokens = model.generate(
39
+ **encoded_ar,
40
+ forced_bos_token_id=tokenizer.lang_code_to_id["en_XX"]
41
+ )
42
+ tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
43
+ # => "The Secretary-General of the United Nations says there is no military solution in Syria."
44
+ ```
45
+
46
+
47
+ See the [model hub](https://huggingface.co/models?filter=mbart-50) to look for more fine-tuned versions.
48
+
49
+
50
+ ## Languages covered
51
+ Arabic (ar_AR), Czech (cs_CZ), German (de_DE), English (en_XX), Spanish (es_XX), Estonian (et_EE), Finnish (fi_FI), French (fr_XX), Gujarati (gu_IN), Hindi (hi_IN), Italian (it_IT), Japanese (ja_XX), Kazakh (kk_KZ), Korean (ko_KR), Lithuanian (lt_LT), Latvian (lv_LV), Burmese (my_MM), Nepali (ne_NP), Dutch (nl_XX), Romanian (ro_RO), Russian (ru_RU), Sinhala (si_LK), Turkish (tr_TR), Vietnamese (vi_VN), Chinese (zh_CN), Afrikaans (af_ZA), Azerbaijani (az_AZ), Bengali (bn_IN), Persian (fa_IR), Hebrew (he_IL), Croatian (hr_HR), Indonesian (id_ID), Georgian (ka_GE), Khmer (km_KH), Macedonian (mk_MK), Malayalam (ml_IN), Mongolian (mn_MN), Marathi (mr_IN), Polish (pl_PL), Pashto (ps_AF), Portuguese (pt_XX), Swedish (sv_SE), Swahili (sw_KE), Tamil (ta_IN), Telugu (te_IN), Thai (th_TH), Tagalog (tl_XX), Ukrainian (uk_UA), Urdu (ur_PK), Xhosa (xh_ZA), Galician (gl_ES), Slovene (sl_SI)
52
+
53
+
54
+ ## BibTeX entry and citation info
55
+ ```
56
+ @article{tang2020multilingual,
57
+ title={Multilingual Translation with Extensible Multilingual Pretraining and Finetuning},
58
+ author={Yuqing Tang and Chau Tran and Xian Li and Peng-Jen Chen and Naman Goyal and Vishrav Chaudhary and Jiatao Gu and Angela Fan},
59
+ year={2020},
60
+ eprint={2008.00401},
61
+ archivePrefix={arXiv},
62
+ primaryClass={cs.CL}
63
+ }
64
+ ```
config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/home/suraj/projects/mbart-50/hf_models/mbart-50-large-many-to-many/",
3
+ "_num_labels": 3,
4
+ "activation_dropout": 0.0,
5
+ "activation_function": "relu",
6
+ "add_bias_logits": false,
7
+ "add_final_layer_norm": true,
8
+ "architectures": [
9
+ "MBartForConditionalGeneration"
10
+ ],
11
+ "attention_dropout": 0.0,
12
+ "bos_token_id": 0,
13
+ "classif_dropout": 0.0,
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": 12,
20
+ "decoder_start_token_id": 2,
21
+ "dropout": 0.1,
22
+ "early_stopping": true,
23
+ "encoder_attention_heads": 16,
24
+ "encoder_ffn_dim": 4096,
25
+ "encoder_layerdrop": 0.0,
26
+ "encoder_layers": 12,
27
+ "eos_token_id": 2,
28
+ "forced_eos_token_id": 2,
29
+ "gradient_checkpointing": false,
30
+ "id2label": {
31
+ "0": "LABEL_0",
32
+ "1": "LABEL_1",
33
+ "2": "LABEL_2"
34
+ },
35
+ "init_std": 0.02,
36
+ "is_encoder_decoder": true,
37
+ "label2id": {
38
+ "LABEL_0": 0,
39
+ "LABEL_1": 1,
40
+ "LABEL_2": 2
41
+ },
42
+ "max_length": 200,
43
+ "max_position_embeddings": 1024,
44
+ "model_type": "mbart",
45
+ "normalize_before": true,
46
+ "normalize_embedding": true,
47
+ "num_beams": 5,
48
+ "num_hidden_layers": 12,
49
+ "output_past": true,
50
+ "pad_token_id": 1,
51
+ "scale_embedding": true,
52
+ "static_position_embeddings": false,
53
+ "transformers_version": "4.4.0.dev0",
54
+ "use_cache": true,
55
+ "vocab_size": 250054
56
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:024ddcc796a33d2e4decd4c1bd5fe90ad295aaba9072edb3796a09ef9b755934
3
+ size 2444714899
sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
3
+ size 5069051
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": "<mask>", "additional_special_tokens": ["ar_AR", "cs_CZ", "de_DE", "en_XX", "es_XX", "et_EE", "fi_FI", "fr_XX", "gu_IN", "hi_IN", "it_IT", "ja_XX", "kk_KZ", "ko_KR", "lt_LT", "lv_LV", "my_MM", "ne_NP", "nl_XX", "ro_RO", "ru_RU", "si_LK", "tr_TR", "vi_VN", "zh_CN", "af_ZA", "az_AZ", "bn_IN", "fa_IR", "he_IL", "hr_HR", "id_ID", "ka_GE", "km_KH", "mk_MK", "ml_IN", "mn_MN", "mr_IN", "pl_PL", "ps_AF", "pt_XX", "sv_SE", "sw_KE", "ta_IN", "te_IN", "th_TH", "tl_XX", "uk_UA", "ur_PK", "xh_ZA", "gl_ES", "sl_SI"]}
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"src_lang": null, "tgt_lang": null, "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "cls_token": "<s>", "pad_token": "<pad>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": "<s>", "tokenizer_file": null, "language_codes": "ML50", "special_tokens_map_file": "/home/suraj/projects/mbart-50/mbart-50/special_tokens_map.json", "name_or_path": "/home/suraj/projects/mbart-50/hf_models/mbart-50-large-many-to-many/"}