Instructions to use nullptr04/CoEvo-RSI-homebench with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nullptr04/CoEvo-RSI-homebench with Transformers:

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

pipe = pipeline("image-text-to-text", model="nullptr04/CoEvo-RSI-homebench")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("nullptr04/CoEvo-RSI-homebench")
model = AutoModelForMultimodalLM.from_pretrained("nullptr04/CoEvo-RSI-homebench", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nullptr04/CoEvo-RSI-homebench with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "nullptr04/CoEvo-RSI-homebench"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "nullptr04/CoEvo-RSI-homebench",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/nullptr04/CoEvo-RSI-homebench

SGLang

How to use nullptr04/CoEvo-RSI-homebench 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 "nullptr04/CoEvo-RSI-homebench" \
    --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": "nullptr04/CoEvo-RSI-homebench",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "nullptr04/CoEvo-RSI-homebench" \
        --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": "nullptr04/CoEvo-RSI-homebench",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use nullptr04/CoEvo-RSI-homebench with Docker Model Runner:
```
docker model run hf.co/nullptr04/CoEvo-RSI-homebench
```

CoEvoX · homebench model-harness pair

Model-harness pair for the homebench domain from CoEvo-RSI: When Harnesses Evolve, Which Models Should Follow?


Domain	`homebench`
Model	repo root (Qwen3.5-9B + full-parameter SFT)
Harness	`harness/`
Paper	arXiv:2610.25078

Reported scores

Smart-home instruction following (HomeBench multi-device, %), evaluated with the harness/ in this repo. VM denotes valid instructions and IM denotes invalid ones.

Split	Succ.	F1
VM (valid multi-device)	75.51	89.23
IM (invalid multi-device)	84.54	91.80

Layout

.            # model weights (config.json, *.safetensors, tokenizer, ...)
harness/     # harness source for this pair

Inference

Serve the model with vLLM. The tool-call and reasoning parsers are required, because the harness talks to the model through tool calls.

huggingface-cli download nullptr04/CoEvo-RSI-homebench --local-dir ckpts/CoEvo-RSI-homebench

CUDA_VISIBLE_DEVICES=0 vllm serve ckpts/CoEvo-RSI-homebench \
  --host 127.0.0.1 --port 8000 \
  --served-model-name CoEvo-RSI-homebench \
  --max-model-len 65536 \
  --enable-auto-tool-choice --tool-call-parser qwen3_coder \
  --reasoning-parser qwen3

Then, from the CoEvoX repository, run the evaluation with the evolved harness (generated, the same as harness/ here). The script reads the model from POLICY_BASE_URL (default http://127.0.0.1:8000/v1).

bash scripts/run_homebench_full_test.sh generated

Citation

If you use this model or harness, please cite:

@misc{su2026coevorsiharnessesevolvemodels,
      title={CoEvo-RSI: When Harnesses Evolve, Which Models Should Follow?},
      author={Jiaqi Su and Guibin Zhang and Jie-Jing Shao and Cong Pang and Tiankuo Yao and Ivor Tsang and Haiyan Yin and Shuicheng Yan},
      year={2026},
      eprint={2610.25078},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2610.25078},
}

Downloads last month: 97

Safetensors

Model size

1.47M params

Tensor type

BF16

Collection including nullptr04/CoEvo-RSI-homebench

CoEvo-RSI

Collection

3 items • Updated 12 days ago • 1

Paper for nullptr04/CoEvo-RSI-homebench

CoEvo-RSI: When Harnesses Evolve, Which Models Should Follow?

Paper • 2610.25078 • Published 10 days ago • 120