hr-persona-bd-llama3.2-3b-gguf

Fine-tuned model for Bangladesh Labour Law and HR practices.

Model Details

• Base Model: unsloth/Llama-3.2-3B-Instruct
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Training Framework: Unsloth
• Model Type: GGUF

Intended Use

This model is designed for:

• Answering questions about Bangladesh Labour Act 2006
• Providing HR guidance for Bangladesh workplaces
• Explaining worker rights and employer obligations
• General HR policy assistance

Usage

With Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("remon-rakibul/hr-persona-bd-llama3.2-3b-gguf")
tokenizer = AutoTokenizer.from_pretrained("remon-rakibul/hr-persona-bd-llama3.2-3b-gguf")

messages = [
    {"role": "user", "content": "What is the maximum working hours per week in Bangladesh?"}
]

inputs = tokenizer.apply_chat_template(messages, return_tensors="pt")
outputs = model.generate(inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0]))

With Unsloth (Faster)

from unsloth import FastLanguageModel

model, tokenizer = FastLanguageModel.from_pretrained("remon-rakibul/hr-persona-bd-llama3.2-3b-gguf")
FastLanguageModel.for_inference(model)

With Ollama (GGUF version)

ollama run hr-persona-bd

Training Details

• Dataset: Bangladesh Labour Act 2006 QA pairs
• Epochs: 1-3
• Learning Rate: 2e-4
• LoRA Rank: 16
• LoRA Alpha: 16

Limitations

• Knowledge is limited to Bangladesh Labour Act 2006 (amended up to 2018)
• Should not be used as sole legal advice - consult legal professionals
• May not cover all edge cases or recent amendments

Citation

@model{hr_persona_bd,
  title={HR Persona Bangladesh},
  year={2026},
  base_model={unsloth/Llama-3.2-3B-Instruct},
  url={https://huggingface.co/remon-rakibul/hr-persona-bd-llama3.2-3b-gguf}
}