Instructions to use AquilaX-AI/security_assistant_2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use AquilaX-AI/security_assistant_2 with Transformers:

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

pipe = pipeline("text-generation", model="AquilaX-AI/security_assistant_2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("AquilaX-AI/security_assistant_2")
model = AutoModelForCausalLM.from_pretrained("AquilaX-AI/security_assistant_2")
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

vLLM

How to use AquilaX-AI/security_assistant_2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "AquilaX-AI/security_assistant_2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AquilaX-AI/security_assistant_2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/AquilaX-AI/security_assistant_2

SGLang

How to use AquilaX-AI/security_assistant_2 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 "AquilaX-AI/security_assistant_2" \
    --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": "AquilaX-AI/security_assistant_2",
		"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 "AquilaX-AI/security_assistant_2" \
        --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": "AquilaX-AI/security_assistant_2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use AquilaX-AI/security_assistant_2 with Docker Model Runner:
```
docker model run hf.co/AquilaX-AI/security_assistant_2
```

INFERENCE

import time
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM, TextStreamer

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
finetuned_model = AutoModelForCausalLM.from_pretrained("AquilaX-AI/security_assistant_2")
tokenizer = AutoTokenizer.from_pretrained("AquilaX-AI/security_assistant")

finetuned_model.to(device)

prompt = """<|im_start|>system
You are a helpful AI assistant named Securitron<|im_end|>
<|im_start|>user
cwe_id:CWE-20
cwe_name:Improper Input Validation
affected_line:Pattern Undefined (v3)
partial_code:example: c4d5ea2f-81a2-4a05-bcd3-202126ae21df
        name:
          type: string
          example: Toolbox
        serial:
file_name:itemit_openapi.yaml
status:True Positive
reason: There is no pattern property that could lead to insufficient input validation.
remediation_action: Always define a pattern to ensure strict input validation.

How to fix this?<|im_end|>
<|im_start|>assistant
"""

s = time.time()

encodeds = tokenizer(prompt, return_tensors="pt",truncation=True).input_ids.to(device)
text_streamer = TextStreamer(tokenizer, skip_prompt = True)

# Increase max_new_tokens if needed
response = finetuned_model.generate(
        input_ids=encodeds,
        streamer=text_streamer,
        max_new_tokens=512,
        use_cache=True,
        pad_token_id=151645,
        eos_token_id=151645,
        num_return_sequences=1
    )
e = time.time()
print(f'time taken:{e-s}')

Downloads last month: 12

Safetensors

Model size

0.5B params

Tensor type

BF16