Update README.md

.gitattributes

@@ -1,4 +1,4 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
+*.jsonl filter=lfs diff=lfs merge=lfs -text*.7z filter=lfs diff=lfs merge=lfs -text
 *.arrow filter=lfs diff=lfs merge=lfs -text
 *.bin filter=lfs diff=lfs merge=lfs -text
 *.bz2 filter=lfs diff=lfs merge=lfs -text

README.md

@@ -2,16 +2,25 @@
 license: apache-2.0
 ---
 
-#  🚀 GroundingSuite - Measuring Complex Multi-Granular Pixel Grounding
-GSEval is a comprehensive benchmark for pixel-level and bounding box-level grounding tasks, designed to evaluate how well AI systems can understand and localize objects or regions in images based on natural language descriptions.
+# 🚀 GSEval - A Comprehensive Grounding Evaluation Benchmark
 
-## 📊 Download GSEval
-   ```
-    git lfs install
-    git clone https://huggingface.co/datasets/hustvl/GSEval
-   ```
+GSEval is a meticulously curated evaluation benchmark consisting of 3,800 images, designed to assess the performance of pixel-level and bounding box-level grounding models. It evaluates how well AI systems can understand and localize objects or regions in images based on natural language descriptions.
 
+## 📊 Download GSEval
+```
+git lfs install
+git clone https://huggingface.co/datasets/hustvl/GSEval
+```
 
 ## 📚 Additional Resources
-- **Paper:** [ArXiv](https://arxiv.org/abs/2503.10596).
-- **GitHub Repository:** [GitHub - GroundingSuite](https://github.com/hustvl/GroundingSuite).
+- **Paper:** [ArXiv](https://arxiv.org/abs/2503.10596)
+- **GitHub Repository:** [GitHub - GroundingSuite](https://github.com/hustvl/GroundingSuite)
+
+### Dataset Preview Configuration
+To preview the dataset content on Hugging Face, add the following configuration to your dataset card:
+
+```yaml
+viewer: true
+```
+
+This configuration enables the dataset preview feature on the Hugging Face hub, allowing users to browse through the JSONL file contents directly on the website.

dataset_infos.json

@@ -0,0 +1,16 @@
+{
+  "default": {
+    "features": {
+      "idx": "int64",
+      "image_path": "string",
+      "class_id": "int64",
+      "label": "string",
+      "caption": "string",
+      "box": "float32[4]",
+      "segmentation": {
+        "size": "int32[2]",
+        "counts": "string"
+      }
+    }
+  }
+}