Datasets:
metadata
license: cc-by-4.0
language:
- en
- ta
📜 Colloquial Tamil Dataset for Machine Translation
This dataset is designed to help AI models translate English text into colloquial Tamil. Unlike standard Tamil, colloquial Tamil includes slang, informal phrases, and regional expressions commonly used in everyday conversations.
📂 Dataset Details
- Languages: English → Colloquial Tamil
- Format: JSONL (
.jsonl
) - Total Entries: 16,269 sentence pairs
- Structure: Each entry contains:
"instruction"
: The task description"input"
: English text"output"
: Colloquial Tamil translation
📜 Dataset Composition
This dataset is a combination of:
1️⃣ jarvisvasu/english-to-colloquial-tamil
- A publicly available dataset for English-to-Colloquial Tamil translations.
2️⃣ YouTube Comments Dataset (Custom-Created)
- Extracted using the YouTube API from selected videos.
- Converted into colloquial Tamil to ensure authenticity and linguistic accuracy.
3️⃣ ChatGPT-Generated Data
- Used to generate additional colloquial Tamil phrases.
- Ensured alignment with natural speech patterns and real-world usage.
This combined dataset was specifically curated to fine-tune the suriya7/English-to-Tamil model, resulting in janisrebekahv/finetuned-colloquial-tamil.
This fine-tuned model enhances English-to-Tamil translations by incorporating natural, conversational, and slang-based Tamil expressions.
🔹 Example Data
{
"instruction": "Translate provided English text into colloquial Tamil",
"input": "Bro, you’ve got to try this dish, it’s amazing!",
"output": "ப்ரோ, இது சாப்பிடிட்டு பாரு, ரொம்ப சூப்பர்!"
}
---