nassimjp/Medic_Chat-Pashto-sft

🩺 Medic_Chat‑Pashto‑sft

A high‑quality Pashto medical question‑answering dataset designed for health‑focused chatbots, clinical reasoning models, and Pashto LLM fine‑tuning.
This dataset contains natural medical queries and safe, helpful assistant responses in clear Pashto.

---

📌 Dataset Summary

Medic_Chat‑Pashto‑sft provides curated medical Q&A pairs in Pashto, formatted in the Ministral‑Instruct chat schema:

{
  "messages": [
    {"role": "user", "content": "..."},
    {"role": "assistant", "content": "..."}
  ]
}

The dataset is suitable for:

• Medical chatbots
• Health‑advice assistants
• Pashto clinical reasoning models
• Safety‑aligned LLM training
• Low‑resource medical NLP research

---

🏷 Tags

language:
  - ps
tags:
  - pashto
  - medical
  - healthcare
  - qa
  - chatbot
  - llm
  - sft
  - instruction
task_categories:
  - text-generation
license: "cc-by-nc-4.0"
pretty_name: "Medic Chat Pashto SFT"

---

📂 Data Structure

Each entry contains:

Field	Description
messages	List of chat turns
role	"user" or "assistant"
content	Pashto text

Example:

{
  "messages": [
    {
      "role": "user",
      "content": "د ژمي د خټکي پوستکي څه ګټې لري؟"
    },
    {
      "role": "assistant",
      "content": "د ژمی د خټکی پوستکی د تودوخی له منځه وړلو او د بدن د میتابولیزم په تنظیم کې مرسته کوي..."
    }
  ]
}

---

🧠 Use Cases

• Pashto medical chatbots
• Health‑advice assistants
• Clinical reasoning SFT
• Medical Q&A generation
• Low‑resource biomedical NLP

---

🛠 Training Compatibility

This dataset is fully compatible with:

• Ministral‑Instruct
• LLaMA‑3 / LLaMA‑2 chat templates
• Qwen chat template
• Unsloth SFT pipelines

Example:

tokenizer.apply_chat_template(
    example["messages"],
    tokenize=True,
    add_generation_prompt=False
)

---

📥 Loading the Dataset

from datasets import load_dataset

ds = load_dataset("nassimjp/Medic_Chat-Pashto-sft")

---

📜 License

This dataset is released under CC‑BY‑NC‑4.0.
It may be used for research and non‑commercial applications.

---

🙌 Acknowledgements

This dataset is part of the broader effort to build high‑quality Pashto medical AI resources and support low‑resource healthcare NLP.

---