metadata

dataset_info:
  features:
    - name: en
      dtype: string
    - name: vi
      dtype: string
  splits:
    - name: test
      num_bytes: 192744
      num_examples: 2000
    - name: train
      num_bytes: 82614470
      num_examples: 1000000
    - name: validation
      num_bytes: 194721
      num_examples: 2000
  download_size: 59201490
  dataset_size: 83001935
task_categories:
  - translation
language:
  - en
  - vi
tags:
  - LM
size_categories:
  - 1M<n<10M

Opus100

Source: https://huggingface.co/datasets/opus100
Num examples:
- 1,000,000 (train)
- 2,000 (validation)
- 192,744 (test)
Language: English

from datasets import load_dataset

load_dataset("tdtunlp/opus100_envi")

Format for Translation task

def preprocess(
    sample,
    instruction_key="### Instruction:",
    input_key="Input:",
    response_key="<|endofprompt|>",
    end_key="<|endoftext|>",
    en2vi=True,
):
    if en2vi:
        if random.random() < 0.5:
            instruction = "Translate the following sentences from English into Vietnamese."
        else:
            instruction = "Dịch các câu sau từ tiếng Anh sang tiếng Việt."
        
        input = sample['en'].strip()
        response = sample['vi'].strip()
    else:
        if random.random() < 0.5:
            instruction = "Translate the following sentences from Vietnamese into English."
        else:
            instruction = "Dịch các câu sau từ tiếng Việt sang tiếng Anh."
        
        input = sample['vi'].strip()
        response = sample['en'].strip()

    return {'text': """Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.
{instruction_key}
{instruction}
{input_key} 
{input}
{response_key}
{response}
{end_key}""".format(
    instruction_key=instruction_key,
    instruction=instruction,
    input_key=input_key,
    input=input,
    response_key=response_key,
    response=response,
    end_key=end_key,
)}

"""
Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.
### Instruction:
Dịch các câu sau từ tiếng Anh sang tiếng Việt.
Input: 
Toast falls jelly-side down, children hit tables and people get hurt.
<|endofprompt|>
Bánh mì nướng rơi đông lại, trẻ con va vào bàn và con người bị thương.
<|endoftext|>
"""