Instructions to use Shubhwtf/ministral-legis-lora with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Shubhwtf/ministral-legis-lora with Transformers:

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

pipe = pipeline("text-generation", model="Shubhwtf/ministral-legis-lora")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Shubhwtf/ministral-legis-lora", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Shubhwtf/ministral-legis-lora with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/Shubhwtf/ministral-legis-lora

SGLang

How to use Shubhwtf/ministral-legis-lora 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 "Shubhwtf/ministral-legis-lora" \
    --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": "Shubhwtf/ministral-legis-lora",
		"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 "Shubhwtf/ministral-legis-lora" \
        --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": "Shubhwtf/ministral-legis-lora",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Shubhwtf/ministral-legis-lora with Docker Model Runner:
```
docker model run hf.co/Shubhwtf/ministral-legis-lora
```

Ministral Legal Reasoning LoRA

A QLoRA fine-tuned legal reasoning adapter trained for grounded California statutory question answering and legal reasoning tasks.

This model was trained to:

answer using only provided statutory excerpts
reduce hallucinations
refuse unsupported claims
provide concise citation-aware responses
perform legal QA over California legislative text

Base Model

Ministral / Mistral reasoning model
Quantized 4-bit QLoRA training
Fine-tuned using Unsloth

Training Objective

The model was supervised on synthetic legal reasoning data generated from California legislative and statutory excerpts.

Training focused heavily on:

grounded statutory reasoning
hallucination resistance
refusal behavior
concise legal answers
section-aware responses
lightweight chain-of-thought reasoning

Dataset Characteristics

The dataset contains:

statutory QA
obligations
timelines
reporting requirements
applicability questions
refusal examples
adversarial hallucination-resistance prompts

Examples of trained refusal behavior:

"The provided text does not specify this information."

Example Usage

Load Base Model + LoRA Adapter

from unsloth import FastLanguageModel
from peft import PeftModel

base_model, tokenizer = FastLanguageModel.from_pretrained(
    "YOUR_BASE_MODEL"
)

model = PeftModel.from_pretrained(
    base_model,
    "Shubhwtf/ministral-legis-lora"
)

Example Prompt

Bill: AB1064

Statutory Excerpt:
Section 22757.21(c)(2)(B) excludes systems designed solely for research or technical assistance.

Question:
Does the law apply to systems used solely for research purposes?

Example Output

Section 22757.21(c)(2)(B) excludes systems designed solely for research or technical assistance from the scope of the law.

Downloads last month: -; Downloads are not tracked for this model. How to track

Model tree for Shubhwtf/ministral-legis-lora

Base model

mistralai/Ministral-8B-Instruct-2410

Adapter

(93)

this model