Add VDEBench evaluation results
Browse filesHi 👋 this PR adds this model's **[VDEBench](https://huggingface.co/datasets/hongzhuyi/VDEBench)** scores to `.eval_results/`, so they show up on the model page and on the VDEBench leaderboard.
VDEBench evaluates instruction-guided editing of dense English/Chinese document pages (3,506 examples). Each score is the equally weighted mean of IoU, CDM, BLEU-4 and TEDS (0–1, higher is better):
| Task | Score |
|---|---:|
| `local` (edited region, headline) | 0.685 |
| `global` (full page) | 0.673 |
Scores come from the VDEBench paper ([arXiv:2602.00122v4](https://arxiv.org/abs/2602.00122v4), Table 2), and the VDEBench authors confirmed both the values and the evaluated checkpoint ([discussion](https://huggingface.co/datasets/hongzhuyi/VDEBench/discussions/2)).
While this PR is open the scores appear as community-provided; merging it makes them official, and closing it removes them. Docs: https://huggingface.co/docs/hub/eval-results · https://huggingface.co/docs/hub/leaderboard-data-guide
Thanks!
- .eval_results/vdebench.yaml +22 -0
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
- dataset:
|
| 2 |
+
id: hongzhuyi/VDEBench
|
| 3 |
+
task_id: local
|
| 4 |
+
value: 0.685
|
| 5 |
+
date: '2026-10-15'
|
| 6 |
+
source:
|
| 7 |
+
url: https://arxiv.org/abs/2602.00122v4
|
| 8 |
+
name: VDEBench paper (v4), Table 2
|
| 9 |
+
user: hongzhuyi
|
| 10 |
+
notes: local setting, n=3,506, PaddleOCR-VL; IoU 0.701, CDM 0.794, BLEU-4 0.472,
|
| 11 |
+
TEDS 0.774
|
| 12 |
+
- dataset:
|
| 13 |
+
id: hongzhuyi/VDEBench
|
| 14 |
+
task_id: global
|
| 15 |
+
value: 0.673
|
| 16 |
+
date: '2026-10-15'
|
| 17 |
+
source:
|
| 18 |
+
url: https://arxiv.org/abs/2602.00122v4
|
| 19 |
+
name: VDEBench paper (v4), Table 2
|
| 20 |
+
user: hongzhuyi
|
| 21 |
+
notes: global setting, n=3,506, PaddleOCR-VL; IoU 0.689, CDM 0.803, BLEU-4 0.430,
|
| 22 |
+
TEDS 0.769
|