Image-Text-to-Text
Safetensors
MLX
mlx-vlm
mistral3
apple-silicon
pixtral
guardrail
content-moderation
safety-classification
multimodal
4-bit precision
conversational
Instructions to use AXONVERTEX-AI-RESEARCH/Shieldstral-1.0-3B-MLX-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use AXONVERTEX-AI-RESEARCH/Shieldstral-1.0-3B-MLX-4bit with MLX:
# Make sure mlx-vlm is installed # pip install --upgrade mlx-vlm from mlx_vlm import load, generate from mlx_vlm.prompt_utils import apply_chat_template from mlx_vlm.utils import load_config # Load the model model, processor = load("AXONVERTEX-AI-RESEARCH/Shieldstral-1.0-3B-MLX-4bit") config = load_config("AXONVERTEX-AI-RESEARCH/Shieldstral-1.0-3B-MLX-4bit") # Prepare input image = ["http://images.cocodataset.org/val2017/000000039769.jpg"] prompt = "Describe this image." # Apply chat template formatted_prompt = apply_chat_template( processor, config, prompt, num_images=1 ) # Generate output output = generate(model, processor, formatted_prompt, image) print(output) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
| #!/usr/bin/env python3 | |
| from __future__ import annotations | |
| import argparse, json, os | |
| from pathlib import Path | |
| from shieldstral_policy import discover_model, score_policy | |
| def main(): | |
| root=Path(__file__).resolve().parents[1] | |
| p=argparse.ArgumentParser(description='Score an AXONVERTEX deployment policy outside the Appendix B taxonomy.') | |
| p.add_argument('--policy',required=True) | |
| p.add_argument('--document',required=True) | |
| p.add_argument('--document-type',choices=('prompt','response'),default='prompt') | |
| p.add_argument('--instruct',default='Evaluate whether the document matches the query criteria.') | |
| p.add_argument('--threshold',type=float,default=0.5) | |
| p.add_argument('--base-url',default=os.getenv('BASE_URL','http://127.0.0.1:18190/v1')) | |
| p.add_argument('--model',default=None) | |
| a=p.parse_args() | |
| registry=json.loads((root/'taxonomy/deployment_policies.json').read_text(encoding='utf-8')) | |
| policies={x['id']:x for x in registry['policies']} | |
| pid=a.policy.upper() | |
| if pid not in policies: raise SystemExit(f'Unknown deployment policy: {pid}') | |
| policy=policies[pid]; query=policy[f'{a.document_type}_query']; model=discover_model(a.base_url,a.model) | |
| score,_=score_policy(base_url=a.base_url,model=model,instruct=a.instruct,query=query,document=a.document,threshold=a.threshold) | |
| print(json.dumps({'classification_mode':'deployment_policy_binary','registry_disclosure':registry['disclosure'],'model':model,'document_type':a.document_type,'policy':{'id':pid,'name':policy['name'],'query':query,**score}},indent=2,ensure_ascii=False)) | |
| if __name__=='__main__': main() | |