Upload README.md with huggingface_hub

README.md

@@ -1,38 +1,48 @@
 ---
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-dataset_info:
-  features:
-  - name: observation.state
-    list: float64
-  - name: action
-    list: float64
-  - name: episode_index
-    dtype: int64
-  - name: frame_index
-    dtype: int64
-  - name: timestamp
-    dtype: float64
-  - name: next.reward
-    dtype: float32
-  - name: next.done
-    dtype: bool
-  - name: next.success
-    dtype: bool
-  - name: index
-    dtype: int64
-  - name: task_index
-    dtype: int64
-  splits:
-  - name: train
-    num_bytes: 88950
-    num_examples: 600
-  download_size: 35684
-  dataset_size: 88950
+# For reference on dataset card metadata, see the spec: https://github.com/huggingface/hub-docs/blob/main/datasetcard.md
+# Doc / guide: https://huggingface.co/docs/hub/datasets-cards
+annotations_creators:
+- no-annotation
+language:
+- en
+license: mit
+multilinguality:
+- monolingual
+size_categories:
+- 1K<n<10K
+source_datasets:
+- original
+task_categories:
+- computer-vision
+task_ids:
+- image-classification
+pretty_name: Sim Bucket-to-Bin Dataset
 ---
-# Sim Bucket-to-Bin Dataset
 
-This dataset contains simulated episodes of a robot pushing a cube.
+This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).
+
+## Dataset Description
+
+Simulated robot control episodes for learning from pixels and interaction.
+
+- **Homepage:** https://github.com/huggingface/lerobot
+- **Paper:** [More Information Needed]
+- **License:** MIT
+
+## Dataset Structure
+
+- Episodes contain RGB camera data and robot state.
+- Data is stored in Parquet files.
+- Video files are MP4 compressed per camera.
+
+## Citation
+
+**BibTeX:**
+
+```bibtex
+@misc{lerobot2024,
+  author = {HuggingFace Robotics Team},
+  title = {LeRobot: Learning from Simulation at Scale},
+  year = {2024},
+  url = {https://github.com/huggingface/lerobot}
+}