metadata
license: apache-2.0
tags:
- qwen2.5
- 14b
- reinforcement-learning
- besiegefield
- catapult
- gemini-2.5-pro
- synthetic-data
- instruct
- transformers
language:
- en
base_model:
- Qwen/Qwen2.5-14B-Instruct
Qwen2.5-14B-Instruct-BesiegeField-CarRL
Qwen2.5-14B-Instruct fine-tuned with Gemini-2.5-Pro synthetic cold-start data and reinforcement-learning optimized for the Car task inside the BesiegeField environment.
📎 Links
- Project Page: https://besiegefield.github.io/
- GitHub: https://github.com/Godheritage/BesiegeField
- arXiv: https://arxiv.org/abs/2510.14980
If you found this model useful, please cite:
@article{zhang2025besiegefield,
title={Agentic Design of Compositional Machines},
author={Zhang, Wenqian and Liu, Weiyang and Liu, Zhen},
journal={arXiv preprint arXiv:2510.14980},
year={2025}
}