|
--- |
|
license: apache-2.0 |
|
datasets: |
|
- internlm/Lean-Workbook |
|
- internlm/Lean-Github |
|
- AI-MO/NuminaMath-CoT |
|
language: |
|
- en |
|
base_model: |
|
- Qwen/Qwen2.5-Math-7B |
|
pipeline_tag: text-generation |
|
library_name: transformers |
|
tags: |
|
- lean4 |
|
- theorem-proving |
|
- formal-mathematics |
|
--- |
|
|
|
<div align="center"> |
|
<h1 style="font-size: 2.0em;">π BFS-Prover: Scalable Best-First Tree Search for LLM-based Automatic Theorem Proving</h1> |
|
<div style="display: flex; justify-content: center; gap: 8px; flex-wrap: wrap;"> |
|
<a href="https://arxiv.org/abs/2502.03438"><img src="https://img.shields.io/badge/arXiv-2502.03438-b31b1b.svg" alt="arXiv"></a> |
|
<a href="https://choosealicense.com/licenses/apache-2.0/"><img src="https://img.shields.io/badge/License-Apache%202.0-blue.svg" alt="License: Apache 2.0"></a> |
|
<a href="https://github.com/leanprover-community/mathlib4"><img src="https://img.shields.io/badge/Lean-4-orange" alt="Lean 4"></a> |
|
</div> |
|
<h2>State-of-the-art tactic generation model in Lean4</h2> |
|
</div> |
|
|
|
This repository contains the latest tactic generator model checkpoint from BFS-Prover, a state-of-the-art theorem proving system in Lean4. While the full BFS-Prover system integrates multiple components for scalable theorem proving, we are releasing the core tactic generation model here. Given a proof state in Lean4, the model generates a tactic that transforms the current proof state into a new state, progressively working towards completing the proof. |
|
|
|
**π Paper: [BFS-Prover: Scalable Best-First Tree Search for LLM-based Automatic Theorem Proving](https://arxiv.org/abs/2502.03438)** |
|
|
|
|
|
## β¨ Model Details |
|
|
|
- Base Model: Qwen2.5-Math-7B |
|
- Training Approach: |
|
- Supervised Fine-Tuning (SFT) on state-tactic pairs |
|
- Direct Preference Optimization (DPO) using compiler feedback |
|
- Training Data Sources: |
|
- Mathlib (via LeanDojo) |
|
- Lean-Github repositories |
|
- Lean-Workbook |
|
- Autoformalized NuminaMath-CoT dataset |
|
|
|
## π Performance |
|
BFS-Prover achieves state-of-the-art performance on the MiniF2F test benchmark. Here's a detailed comparison: |
|
|
|
### π MiniF2F Test Benchmark Results |
|
|
|
| Prover System | Search Method | Critic Model | Tactic Budget | Score | |
|
|---------------|---------------|--------------|---------------|--------| |
|
| BFS-Prover | BFS | No | Accumulative | **72.95%** | |
|
| BFS-Prover | BFS | No | 2048Γ2Γ600 | **70.83% Β± 0.89%** | |
|
| HunyuanProver | BFS | Yes | 600Γ8Γ400 | 68.4% | |
|
| InternLM2.5-StepProver | BFS | Yes | 256Γ32Γ600 | 65.9% | |
|
| DeepSeek-Prover-V1.5 | MCTS | No | 32Γ16Γ400 | 63.5% | |
|
|
|
|
|
### π Key Advantages |
|
- β
Achieves better performance without requiring a critic model (value function) |
|
- β
Combined with simpler search method (BFS) rather than MCTS |
|
|
|
## βοΈ Usage |
|
- The model expects Lean4 tactic states in the format `"{state}:::"` |
|
- `:::` serves as a special indicator to signal the model to generate a tactic for the given state. |
|
- The model will echo back the input state followed by the generated tactic. |
|
|
|
|
|
```python |
|
# Example code for loading and using the tactic generator model |
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
|
model = AutoModelForCausalLM.from_pretrained("bytedance-research/BFS-Prover") |
|
tokenizer = AutoTokenizer.from_pretrained("bytedance-research/BFS-Prover") |
|
state = "h : x = y + 2 β’ x - 1 = y + 1" |
|
sep = ":::" |
|
prompt = state + sep # Creates "h : x = y + 2 β’ x - 1 = y + 1:::" |
|
|
|
inputs = tokenizer(prompt, return_tensors="pt") |
|
outputs = model.generate(**inputs) |
|
tactic = tokenizer.decode(outputs[0], skip_special_tokens=True).split(sep)[1] |
|
print(tactic) |
|
|
|
# Complete example: |
|
# Input state: "h : x = y + 2 β’ x - 1 = y + 1" |
|
# Full prompt: "h : x = y + 2 β’ x - 1 = y + 1:::" |
|
# Model output: "h : x = y + 2 β’ x - 1 = y + 1:::simp [h]" |
|
# Final tactic: "simp [h]" |
|
``` |
|
|
|
## π Citation |
|
|
|
If you use this model in your research, please cite our paper: |
|
|
|
```bibtex |
|
@article{xin2025bfs, |
|
title={BFS-Prover: Scalable Best-First Tree Search for LLM-based Automatic Theorem Proving}, |
|
author={Xin, Ran and Xi, Chenguang and Yang, Jie and Chen, Feng and Wu, Hang and Xiao, Xia and Sun, Yifan and Zheng, Shen and Shen, Kai}, |
|
journal={arXiv preprint arXiv:2502.03438}, |
|
year={2025} |
|
} |
|
``` |
|
|
|
## π License |
|
|
|
https://choosealicense.com/licenses/apache-2.0/ |
|
|
|
## π§ Contact |
|
|
|
For questions and feedback about the tactic generator model, please contact: |
|
- Ran Xin ([email protected]) |
|
- Kai Shen ([email protected]) |