Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -34,6 +34,8 @@ Cryo-Bench includes five benchmark tasks covering key components of the cryosphe
|
|
| 34 |
|
| 35 |
---
|
| 36 |
|
|
|
|
|
|
|
| 37 |
## 🏆 Benchmark Results
|
| 38 |
|
| 39 |
Table below reports mIoU (↑) for all models evaluated with **frozen encoders** and **100% training data** using the UPerNet decoder. Rank (↓) is averaged across all five tasks. Baseline models (U-Net, ViT) are trained from scratch.
|
|
|
|
| 34 |
|
| 35 |
---
|
| 36 |
|
| 37 |
+
The dataset contains the exact training, validation, and test splits used in **Cryo-Bench**, covering the **SICD, GLID, GLD, GSDD, and CaFFe** datasets.
|
| 38 |
+
|
| 39 |
## 🏆 Benchmark Results
|
| 40 |
|
| 41 |
Table below reports mIoU (↑) for all models evaluated with **frozen encoders** and **100% training data** using the UPerNet decoder. Rank (↓) is averaged across all five tasks. Baseline models (U-Net, ViT) are trained from scratch.
|