Instructions to use Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1")
model = AutoModelForCausalLM.from_pretrained("Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1

SGLang

How to use Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1 with Docker Model Runner:
```
docker model run hf.co/Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1
```

LegalVision: Sri Lankan Property Law Reasoning Model

Fine-tuned Llama 3.1 8B for Sri Lankan property law with step-by-step legal reasoning, IRAC analysis, and statute citations.

Project Info

Field	Details
Project	LegalVision (ID: 25-26J-127)
Component	Explainable Legal Reasoning Module
Author	S. Sivanuja
Collaborator	Sharan (Knowledge Graph)
Supervisors	Ms. Vindhya Kalapuge, Ms. Adya Dissanayake
Institution	IIT Sri Lanka / University of Westminster

Training Config

Spec	Value
Base Model	Llama-3.1-8B-Instruct
Method	QLoRA (4-bit)
LoRA Rank	64
Epochs	3
Learning Rate	2e-4

Topics: Property Transfer, Bim Saviya, Prescription, Partition, Mortgage, Lease, State Land, Foreign Ownership, Inheritance

Usage

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

bnb_config = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_compute_dtype=torch.bfloat16)
base = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-3.1-8B-Instruct", quantization_config=bnb_config, device_map="auto")
model = PeftModel.from_pretrained(base, "YOUR_USERNAME/legalvision-srilanka-llama3.1-8b")
tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-3.1-8B-Instruct")

messages = [
    {"role": "system", "content": "You are a Sri Lankan property law expert. Provide IRAC analysis with statute citations."},
    {"role": "user", "content": "What are the requirements for property transfer?"}
]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
output = model.generate(**inputs, max_new_tokens=1024, temperature=0.7)
print(tokenizer.decode(output[0], skip_special_tokens=True))

API Deployment (Inference Endpoints)

Includes handler.py for HuggingFace Inference Endpoints.

Request:

{
    "messages": [{"role": "user", "content": "Can foreigners buy land in Sri Lanka?"}],
    "parameters": {"max_new_tokens": 1024, "temperature": 0.7}
}

Response:

{
    "generated_text": "**Answer:** Foreigners face restrictions under the Land (Restrictions on Alienation) Act...",
    "usage": {"prompt_tokens": 156, "completion_tokens": 423, "total_tokens": 579}
}

Limitations

Sri Lankan property law only - not for other jurisdictions
Informational purposes - not legal advice
Consult a qualified lawyer for actual legal matters

Citation

@misc{legalvision2025,
  author = {Sivanuja, S.},
  title = {LegalVision: Sri Lankan Property Law Reasoning Model},
  year = {2025},
  publisher = {HuggingFace},
  note = {Project ID: 25-26J-127, IIT Sri Lanka}
}

Downloads last month: 5

Model tree for Sivanuja/Legal_vision_llama-3.1-8B-instruct-v1

Base model

meta-llama/Llama-3.1-8B

Finetuned

meta-llama/Llama-3.1-8B-Instruct

Finetuned

(2769)

this model