Instructions to use decaf-usenix/Decaf-Gen-1.3b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use decaf-usenix/Decaf-Gen-1.3b with Transformers:

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

pipe = pipeline("text-generation", model="decaf-usenix/Decaf-Gen-1.3b")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("decaf-usenix/Decaf-Gen-1.3b")
model = AutoModelForCausalLM.from_pretrained("decaf-usenix/Decaf-Gen-1.3b")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use decaf-usenix/Decaf-Gen-1.3b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "decaf-usenix/Decaf-Gen-1.3b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "decaf-usenix/Decaf-Gen-1.3b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/decaf-usenix/Decaf-Gen-1.3b

SGLang

How to use decaf-usenix/Decaf-Gen-1.3b 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 "decaf-usenix/Decaf-Gen-1.3b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "decaf-usenix/Decaf-Gen-1.3b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "decaf-usenix/Decaf-Gen-1.3b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "decaf-usenix/Decaf-Gen-1.3b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use decaf-usenix/Decaf-Gen-1.3b with Docker Model Runner:
```
docker model run hf.co/decaf-usenix/Decaf-Gen-1.3b
```

Decaf-Gen-1.3b

Fine-tuned LLM for binary decompilation, part of the Decaf pipeline.

This is the Decaf-Gen generator that produces candidate decompilations from Ghidra-decompiled source. It was fine-tuned on top of LLM4Binary/llm4decompile-1.3b-v1.6 on a merged stripped/unstripped corpus and stopped on the stripped half (the *_stripped_stop checkpoint family from the Decaf paper).

Use

from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained("decaf-usenix/Decaf-Gen-1.3b")
tokenizer = AutoTokenizer.from_pretrained("decaf-usenix/Decaf-Gen-1.3b")

Or via vLLM in the Decaf pipeline; see scripts/inference.py and the configs/post_rebuttal_experiments/ours_base_1.3b/ configs in the Decaf release repo.

Companion repos

Repo	Role
`decaf-usenix/Decaf-Gen-1.3b` / `-6.7b` / `-22b`	Generator at three scales
`decaf-usenix/Decaf-ReRanker-32b-stripped`	Reranker (stripped)
`decaf-usenix/Decaf-ReRanker-32b-unstripped`	Reranker (unstripped)
`decaf-usenix/Decaf-Test-Sets`	ExeBench evaluation sets
`decaf-usenix/Decaf-Juliet-Funceval`	Juliet vulnerability-detection eval inputs

Run scripts/download.sh --all from the release repo to fetch everything.

Downloads last month: 22

Safetensors

Model size

0.2B params

Tensor type

F32

Model tree for decaf-usenix/Decaf-Gen-1.3b

Base model

LLM4Binary/llm4decompile-1.3b-v1.6

Finetuned

(1)

this model