README.md

---
base_model:
- Qwen/Qwen2.5-Coder-7B-Instruct
library_name: transformers
license: mit
metrics:
- accuracy
pipeline_tag: text-generation
---

<div align="center">
  <h1 align="center">
Z1: Efficient Test-time Scaling with Code
</h1>
<p>Train Large Language Model to Reason with Shifted Thinking
</p>
</div>
<p align="center">
  <a href="https://arxiv.org/abs/2504.00810"><b>[📜 Paper]</b></a> •
  <a href="https://huggingface.co/efficientscaling/Z1-7B"><b>[🤗 HF Models]</b></a> •
  <a href="https://github.com/efficientscaling/Z1"><b>[🐱 GitHub]</b></a>
  <!-- <a href="https://9557c5365a6f44dc84.gradio.live"><b>[🐯 Gradio Demo]</b></a> -->
  <br>

  <!-- <a href="#-quick-start">Quick Start</a> • -->
  <!-- <a href="#%EF%B8%8F-citation">Citation</a> -->
</p>



## Model Details
To begin with the shifted thinking mode, please refer to https://github.com/efficientscaling/Z1.

## Evaluation

<p align="left">
    <img src="tts.png" width="800">
    <br>
    <!-- <em>Test-time scaling comparison between Z1-7B and R1-Distill-Qwen-7B. </em> -->
</p>

<!-- ## Example

<p align="center">
    <img src="simple_reason.png" width="1200">
    <br>
    <em>Simple Reasoning</em>
    <br>
    <img src="complex_reason.png" width="1200">
    <br>
    <em>Complex Reasoning</em>
    <br>

</p> -->

## Gradio Demo

```python
import copy
from typing import List
from dataclasses import dataclass

import gradio as gr
from vllm import LLM, SamplingParams
from transformers import  AutoTokenizer

BOX=r"\boxed{}"
ANSWER_WITH_BOX=f"\n\nI overthought it, the final answer in {BOX} should be:\n\n"
ANSWER_WITHOUT_BOX=f"\n\nI overthought it, the final answer should be:\n\n"

model_name = "efficientscaling/Z1-7B"

@dataclass
class ThinkingLLM(LLM):

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)

    def thinking_generate(self, prompts: List[str], sampling_params: SamplingParams = None, max_tokens_for_thinking: int = None):

        # If no SamplingParams is provided, create a default one
        if sampling_params is None:
            raise ValueError("Sampling_params can't be None!")
        else:
            all_max_tokens = sampling_params.max_tokens
            # Override the max_tokens in the provided SamplingParams with the budget
            sampling_params.max_tokens = max_tokens_for_thinking
            print(f"All tokens: {all_max_tokens}")
            print(f"Tokens for thinking: {max_tokens_for_thinking}")

        trajectories = self.generate(prompts, sampling_params)

        rethinking_str = ANSWER_WITHOUT_BOX
        sampling_params.max_tokens = all_max_tokens

        answers = copy.deepcopy(trajectories)

        unfinished_id = []
        thinking_token = 0
        new_prompts = []

        for id, traj in enumerate(trajectories):
            if traj.outputs[0].finish_reason == 'length':
                unfinished_id.append(id)
                new_prompts.append(prompts[id] + traj.outputs[0].text + rethinking_str)
            thinking_token += len(traj.outputs[0].token_ids)

        avg_thinking_token = thinking_token / len(prompts)

        if new_prompts:
            print(new_prompts[0])

            o = self.generate(
                new_prompts,
                sampling_params=sampling_params,
            )
            
        for i, uid in enumerate(unfinished_id):
            answers[uid] = o[i]

        return new_prompts, answers


def generate_text(prompt, max_tokens, max_tokens_for_thinking, temperature, top_p):

    sampling_params = SamplingParams(
        temperature=temperature,
        max_tokens=max_tokens,
        top_p=top_p,
        skip_special_tokens=False,
    )

    trajectories, outputs = llm.thinking_generate(prompt, sampling_params, max_tokens_for_thinking=max_tokens_for_thinking)
    return trajectories[0] + '\n\n' + outputs[0].outputs[0].text if trajectories else outputs[0].outputs[0].text


llm = ThinkingLLM(
    model=model_name,
    tensor_parallel_size=1,
    gpu_memory_utilization=0.96,
    )


with gr.Blocks() as demo:
    gr.Markdown("# Reason with shifted thinking")
    
    with gr.Row():
        with gr.Column():
            prompt_input = gr.Textbox(
                label="Prompt",
                placeholder="Input",
                lines=5,
            )
            max_tokens_for_thinking_input = gr.Slider(
                label="shifted_thinking_window_size",
                minimum=1,
                maximum=32786,
                value=4000,
                step=1,
            )
            max_tokens_input = gr.Slider(
                label="all_max_tokens",
                minimum=1,
                maximum=32786,
                value=32786,
                step=1,
            )
            temperature_input = gr.Slider(
                label="Temperature",
                minimum=00,
                maximum=2.0,
                value=0,
                step=0.1,
            )
            top_p_input = gr.Slider(
                label="Top-p",
                minimum=0.0,
                maximum=1.0,
                value=1,
                step=0.01,
            )
            generate_button = gr.Button("Generate")
        
        with gr.Column():
            output_text = gr.Textbox(
                label="Shifted Thinking Window",
                placeholder="Text is here...",
                lines=10,
            )
    
    generate_button.click(
        fn=generate_text,
        inputs=[prompt_input, max_tokens_for_thinking_input,max_tokens_input,  temperature_input, top_p_input],
        outputs=output_text, 
    )

if __name__ == "__main__":
    demo.launch()
```

## Citation
```
@misc{yu2025efficientscaling,
      title={Z1: Efficient Test-time Scaling with Code}, 
      author={Zhaojian Yu and Yinghao Wu and Yilun Zhao and Arman Cohan and Xiao-Ping Zhang},
      year={2025},
      eprint={2504.00810},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2504.00810}, 
}
```