Edit model card
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/model-cards#model-card-metadata)

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

!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
Downloads last month
2
Safetensors
Model size
72.2M params
Tensor type
F32
ยท
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Spaces using wolfrage89/annual_report_translation_id_en 2