| base_model: | |
| - Qwen/Qwen3-8B | |
| datasets: | |
| - Heng1999/dapo-en-10k | |
| - Heng1999/Omni-MATH-512 | |
| language: | |
| - en | |
| license: apache-2.0 | |
| tags: | |
| - tool-use | |
| - reasoning | |
| - mathematics | |
| - qwen | |
| pipeline_tag: text-generation | |
| library_name: transformers | |
| paper: 2508.19201 | |
| # Understanding Tool-Integrated Reasoning (TIR Model) | |
| This repository contains the official DAPO-trained Tool-Integrated Reasoning (TIR) model from the paper: **[Understanding Tool-Integrated Reasoning](https://arxiv.org/abs/2508.19201)**. | |
| This model is a Qwen3-8B fine-tuned using the DAPO algorithm to effectively solve complex mathematical problems by integrating a Python code interpreter. It serves as the primary experimental model in our paper. |