nandhinivaradharajan14/tamil-english-colloquial-translations

Dataset: English-Tamil (en-ta) Parallel Corpus

This dataset contains parallel sentences in English and Tamil (en-ta) that have been curated from multiple sources. It is designed for tasks such as machine translation, language modeling, and other natural language processing (NLP) applications involving English and Tamil.

Dataset Composition

The dataset is composed of three main parts, which have been concatenated into a single file with two columns: ta (Tamil) and en (English).

-Hugging Face Dataset( Sampuran01/eng_to_tamil )

A portion of the dataset is sourced from the Hugging Face dataset repository, specifically the "Sampuran01/eng_to_tamil dataset". This provides an initial set of English-Tamil sentence pairs.

-Tatoeba Sentence Pairs

Additional sentence pairs were downloaded from the https://tatoeba.org/en/downloads , which is a community-driven collection of translated sentences. These pairs were filtered to include only English-Tamil (en-ta) translations.

-YouTube Comments Retrieval and Translation

A significant portion of the dataset was created by retrieving Tamil comments from YouTube. These comments were then translated into English using the Google Cloud Translation API to ensure high-quality translations. This step ensures the dataset includes modern, colloquial, and diverse language usage.

Dataset Structure

The final dataset is a single file in a tabular format with two columns:

ta: The Tamil text.

en: The corresponding English translation.

Example:

ta (Tamil) - en (English)

நல்வரவு - Welcome

இது ஒரு மாதிரி தரவு. - This is a sample dataset.

How the Dataset Was Created

1.Data Collection:

The Sampuran01/eng_to_tamil dataset was downloaded from Hugging Face.

English-Tamil sentence pairs were extracted from the Tatoeba site.

Tamil comments were retrieved from YouTube and translated into English using the Google Cloud Translation API.

2.Data Concatenation:

All three sources were combined into a single dataset, ensuring no duplicates and maintaining alignment between the Tamil and English sentences.

3.Final Format:

The dataset was saved in a tabular format with two columns: ta (Tamil) and en (English).

Intended Use

This dataset is intended for:

Training and evaluating machine translation models (e.g., English to Tamil or Tamil to English). Building language models for Tamil and English. Research in other NLP tasks.

Limitations

The dataset may contain noise, especially in the YouTube comments section, as they were user-generated and may include informal language, typos, or errors. The translations from Google Cloud API, while generally accurate, may not always capture nuanced or context-specific meanings.

How to Download the Dataset?

1.Load the dataset - dataset = load_dataset("nandhinivaradharajan14/tamil-english-colloquial-translations")

2.Convert to Pandas DataFrame - df = dataset["train"].to_pandas()

3.Display the first few rows - print(df.head())

If the dataset is gated (private), authenticate first:

```python
from huggingface_hub import notebook_login

# Log in to Hugging Face
notebook_login()
```

Then reload the dataset with authentication:

    ```python
    dataset = load_dataset("nandhinivaradharajan14/sample", use_auth_token=True)
    ```