from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

# Load model
base_model = AutoModelForCausalLM.from_pretrained(
    "meta-llama/Llama-3.1-8B-Instruct",
    torch_dtype=torch.float16,
    device_map="auto"
)
model = PeftModel.from_pretrained(base_model, "Christine-HiAiPerf/llama-3.1-8b-canadian-tax")
tokenizer = AutoTokenizer.from_pretrained("Christine-HiAiPerf/llama-3.1-8b-canadian-tax")

# Generate response
question = "What is the basic personal amount in Canada?"
prompt = f"<s>[INST] {question} [/INST]"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

outputs = model.generate(
    **inputs,
    max_new_tokens=256,
    temperature=0.7,
    top_p=0.9,
    do_sample=True
)

answer = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(answer.split("[/INST]")[1].strip())

Limitations

Base Model: Llama-3.1-8B-Instruct
Training Method: LoRA (Low-Rank Adaptation)
Domain: Canadian Tax Law (Income Tax Act, Excise Tax Act)
Task: Question Answering / Legal Consulting
Language: English

The model is trained specifically on Canadian tax law and may not perform well on other legal domains
Responses should be verified by qualified tax professionals
The model reflects tax law as of the training data cutoff (February 2026)
Not a substitute for professional legal or tax advice

Ethical Considerations

This model should only be used for educational and informational purposes
Users should not rely solely on model outputs for tax filing or legal decisions
Always consult with qualified tax professionals for specific situations

Citation

bibtex

@misc{llama31-8b-canadian-tax,
  title={Llama-3.1-8B Fine-tuned on Canadian Tax Law},
  author={Fine-tuned from Meta's Llama-3.1-8B-Instruct},
  year={2026},
  url={https://huggingface.co/Christine-HiAiPerf/llama-3.1-8b-canadian-tax}
}

Model Card Contact

For questions or feedback about this model, please open an issue on the model repository.