Instructions to use JetBrains-Research/PIPer-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use JetBrains-Research/PIPer-8B with Transformers:

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

pipe = pipeline("text-generation", model="JetBrains-Research/PIPer-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("JetBrains-Research/PIPer-8B")
model = AutoModelForCausalLM.from_pretrained("JetBrains-Research/PIPer-8B")
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 Settings

vLLM

How to use JetBrains-Research/PIPer-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "JetBrains-Research/PIPer-8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "JetBrains-Research/PIPer-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/JetBrains-Research/PIPer-8B

SGLang

How to use JetBrains-Research/PIPer-8B 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 "JetBrains-Research/PIPer-8B" \
    --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": "JetBrains-Research/PIPer-8B",
		"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 "JetBrains-Research/PIPer-8B" \
        --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": "JetBrains-Research/PIPer-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use JetBrains-Research/PIPer-8B with Docker Model Runner:
```
docker model run hf.co/JetBrains-Research/PIPer-8B
```

nielsr HF Staff commited on Oct 1, 2025

Commit

dce7d27

verified ·

1 Parent(s): 8814b38

Improve model card: Add paper and code links

Browse files

This PR enhances the model card by adding prominent links to the paper ([PIPer: On-Device Environment Setup via Online Reinforcement Learning](https://huggingface.co/papers/2509.25455)) and its GitHub repository ([https://github.com/JetBrains-Research/PIPer](https://github.com/JetBrains-Research/PIPer)).

These additions improve the discoverability and accessibility of key resources for the model. The existing metadata is correct and remains unchanged. No sample usage section has been added as no direct inference code was found in the provided GitHub README.

Files changed (1) hide show

README.md +8 -5

README.md CHANGED Viewed

@@ -1,19 +1,22 @@
 ---
-library_name: transformers
 datasets:
 - JetBrains-Research/PIPer-envbench-zeroshot-rl
 - JetBrains-Research/PIPer-SFT-2500-sharegpt
-base_model:
-- JetBrains-Research/Qwen3-8B-am
-pipeline_tag: text-generation
 license: mit
 ---
 <img src="https://github.com/JetBrains-Research/PIPer/blob/main/misc/piper-logo.png?raw=true" alt="PIPer Mascot" style="height: 6em">
 <h1>
   PIPer: On-Device Environment Setup via Online Reinforcement Learning
 </h1>
 <div align="center">
 [![Models](https://img.shields.io/badge/🤗%20Hugging%20Face-Models-orange.svg)](https://jb.gg/PIPer)
@@ -97,4 +100,4 @@ uv run piper/hparams_entrypoint.py +experiment=llm-reward --info config
 ## 📄 License
-This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.

 ---
+base_model:
+- JetBrains-Research/Qwen3-8B-am
 datasets:
 - JetBrains-Research/PIPer-envbench-zeroshot-rl
 - JetBrains-Research/PIPer-SFT-2500-sharegpt
+library_name: transformers
 license: mit
+pipeline_tag: text-generation
 ---
 <img src="https://github.com/JetBrains-Research/PIPer/blob/main/misc/piper-logo.png?raw=true" alt="PIPer Mascot" style="height: 6em">
 <h1>
   PIPer: On-Device Environment Setup via Online Reinforcement Learning
 </h1>
+[Paper](https://huggingface.co/papers/2509.25455) | [Code](https://github.com/JetBrains-Research/PIPer)
 <div align="center">
 [![Models](https://img.shields.io/badge/🤗%20Hugging%20Face-Models-orange.svg)](https://jb.gg/PIPer)
 ## 📄 License
+This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.