wizardoftrap/Llama-3.2-1B-Indian-history

Indian History LLM – LLaMA 3.2 1B Fine-tuned

• Model name: wizardoftrap/Llama-3.2-1B-Indian-history
• Base model: unsloth/Llama-3.2-1B-Instruct
• Fine-tuning method: LoRA with Unsloth
• Domain: Indian History
• Author: Shiv Prakash (wizardoftrap)

---

Overview

• Indian History LLM is a domain-adapted version of LLaMA 3.2 1B Instruct, fine-tuned to answer questions on Indian History.
• The model is optimized to behave like a history tutor, producing concise, exam-style answers aligned with Indian curricula.
• This model was trained 2x faster with Unsloth and Huggingface's TRL library.
• This model was fine tuned on Google Colab with L4 GPU

---

Base Model

• Model: Meta LLaMA 3.2 1B Instruct
• Architecture: Decoder-only Transformer
• Context length: up to 2048 tokens
• License: As per base model (Meta LLaMA license)

---

OpenVino

• This model is converted to OpenVino IR format for deployement on Intel GPUs Link

Training Dataset

Fine-tuned on: wizardoftrap/indianHistory

• ~2.5K high-quality Q&A pairs
• Instruction-style format
• Focused on Indian history during the colonial period and freedom struggle

Dataset example:

{
  "input": "When did Mahatma Gandhi return to India and what did he advocate?",
  "output": "Gandhi returned in January 1915 and promoted satyagraha, a method of non-violent resistance based on truth and moral force."
}