Modelpublic

StephenChou/cedar-qwen27b-sft-v2

sourceHugging Faceapache-2.0updated 2d agoView on Hugging Face

0likes12downloads

Model Card

Cedar Qwen3.5-27B SFT v2

This is the 27B supervised fine-tuned base model used by the Cedar/RAISE access-control policy generation experiments. It maps natural-language access-control requirements and schemas to Cedar policies.

The model is based on Qwen/Qwen3.5-27B and is the required base for StephenChou/cedar-raise-qwen3.5-27b.

Loading

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "StephenChou/cedar-qwen27b-sft-v2"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
    trust_remote_code=True,
)

For vLLM:

python

from vllm import LLM

llm = LLM(
    model="StephenChou/cedar-qwen27b-sft-v2",
    dtype="bfloat16",
    trust_remote_code=True,
)

Intended use

Research on translating natural-language access-control requirements into Cedar policies. Outputs should be validated with the Cedar parser and semantic checks before deployment.

License

Apache-2.0, following the base model.