Research map / Reinforcement learning / RL from feedback
SupGRPO: Enhancing GRPO with Matching-based Online SFT for Text Spotting
Xudong Xie, Yuzhe Li, Jing Shi, Zhifei Zhang, Curtis Wigington, Zhaowen Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add ATS data to training (ablation) | SupGRPO 3B · ATS detection | detection score: 65.3 -> 71.6 | Improved |
| add ATS data to training (ablation) | SupGRPO 3B · ATS recognition | recognition score: 75.7 -> 84.4 | Improved |
| add GRPO joint training to SFT (SupGRPO vs SFT) (ablation) | Qwen2.5-VL-3B · ATS (Qwen2.5-VL-3B) detection | detection score: 63.3 -> 71.6 | Improved |
| add SFT joint training to GRPO (SupGRPO vs GRPO) (ablation) | Qwen2.5-VL-3B · ATS (Qwen2.5-VL-3B) recognition | recognition score: 80.0 -> 84.4 | Improved |
| use GRPO instead of SFT (ablation) | Qwen2.5-VL-3B · ATS (Qwen2.5-VL-3B) detection | detection score: 63.3 -> 62.0 | Hurt |
| add data augmentation to SFT (ablation) | SFT · ATS (Qwen2.5-VL-3B) | detection score: 63.3 -> 65.1 | Improved |
| apply online SFT to all tokens on top of GRPO (ablation) | GRPO · ATS (Qwen2.5-VL-3B) detection | detection score: 62.0 -> 64.0 | Improved |
| apply online SFT to text tokens only on top of GRPO (ablation) | GRPO · ATS (Qwen2.5-VL-3B) detection | detection score: 62.0 -> 60.8 | Hurt |
| apply online SFT to location tokens with matching on top of GRPO (ablation) | GRPO · ATS (Qwen2.5-VL-3B) detection | detection score: 62.0 -> 71.6 | Improved |
| use IoU and text matching instead of IoU-only matching (ablation) | SupGRPO location SFT · ATS (Qwen2.5-VL-3B) detection | detection score: 66.5 -> 71.6 | Improved |
| use IoU and text matching instead of text-only matching (ablation) | SupGRPO location SFT · ATS (Qwen2.5-VL-3B) detection | detection score: 68.6 -> 71.6 | Improved |
| add text reward to baseline model with GRPO (ablation) | Qwen2.5-VL-3B · ATS (Qwen2.5-VL-3B) detection | detection score: 26.3 -> 61.1 | Improved |
Closest papers by meaning
- GoMatching: A Simple Baseline for Video Text Spotting via Long and Short Term Matching (NeurIPS 2024 · similarity 0.87)
- SPTS v2: Single-Point Scene Text Spotting (TPAMI 2023 · similarity 0.87)
- SemiETS: Integrating Spatial and Content Consistencies for Semi-Supervised End-to-end Text Spotting (CVPR 2025 · similarity 0.86)
- Bridging the Gap Between End-to-End and Two-Step Text Spotting (CVPR 2024 · similarity 0.86)
- ODM: A Text-Image Further Alignment Pre-training Approach for Scene Text Detection and Spotting (CVPR 2024 · similarity 0.85)
- Arbitrary Reading Order Scene Text Spotter with Local Semantics Guidance (AAAI 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
