This repository contains the official training and evaluation code for QGR. This anonymous repository provides the inference code for review.
The full model code, training code, checkpoints/configurations, and reproducibility scripts will be released after acceptance.
This repository includes code for evaluation.
1QGR/
2│
3├── src/
4│ ├── infer_QGR_med4.py
5│ ├── infer_QGR_SLAKE.py
6│ ├── infer_Qwen3_med4.py
7│ ├── infer_Qwen3_SLAKE.py
8│ ├── infer_MedGemma_med4.py
9│ └── infer_MedGemma_SLAKE.py
10│
11├── scripts/
12│ ├── run_eval_QGR_llm_judge.sh
13│ ├── run_eval_QGR_SLAKE_llm_judge.sh
14│ ├── run_eval_Qwen3_llm_judge.sh
15│ ├── run_eval_Qwen3_SLAKE_llm_judge.sh
16│ ├── run_eval_MedGemma_llm_judge.sh
17│ └── run_eval_MedGemma_SLAKE_llm_judge.sh
18│
19└── README.md
