| ### Finetuned on annual report sentence pair | |
| This marianMT has been further finetuned on annual report sentence pairs | |
| ## Test out at huggingface spaces! | |
| https://huggingface.co/spaces/wolfrage89/finance_domain_translation_marianMT | |
| ## Sample colab notebook | |
| https://colab.research.google.com/drive/1H57vwiah7n1JXvXYMqJ8dklrIuU6Cljb?usp=sharing | |
| ## How to use | |
| ```python | |
| !pip install transformers | |
| !pip install sentencepiece | |
| from transformers import MarianMTModel, MarianTokenizer | |
| tokenizer = MarianTokenizer.from_pretrained("wolfrage89/annual_report_translation_id_en") | |
| model = MarianMTModel.from_pretrained("wolfrage89/annual_report_translation_id_en") | |
| #tokenizing bahasa sentence | |
| bahasa_sentence = "Interpretasi ini merupakan interpretasi atas PSAK 46: Pajak Penghasilan yang bertujuan untuk mengklarifikasi dan memberikan panduan dalam merefleksikan ketidakpastian perlakuan pajak penghasilan dalam laporan keuangan." | |
| tokenized_bahasa_sentence = tokenizer([bahasa_sentence], return_tensors='pt', max_length=104, truncation=True) | |
| #feeding tokenized sentence into model, the max_legnth have been set to 104 as the model was trained mostly on sentences with this length | |
| translated_tokens = model.generate(**tokenized_bahasa_sentence, max_length=104)[0] | |
| ## decoding the tokens to get english sentence | |
| english_sentence = tokenizer.decode(translated_tokens, skip_special_tokens=True) | |
| print(english_sentence) | |
| # This interpretation is an interpretation of PSAK 46: Income Tax that aims to clarify and provide guidance in reflecting the uncertainty of income tax treatments in the financial statements. | |
| ``` | |
| ### opus-mt-id-en (original model) | |
| * source languages: id | |
| * target languages: en | |
| * OPUS readme: [id-en](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/id-en/README.md) | |