Update README.md

README.md

@@ -4,42 +4,66 @@ size_categories:
 
 # Dataset Card for PRTS Post-Training Data
 
-This dataset card describes the **PRTS Post-Training Data**, which includes high-quality data for robotic manipulation tasks. The dataset covers both **dual-arm** and **single-arm** manipulation, including **fine manipulation** and **long-range operations**. It is curated in the **LeRobot format** to assist in the development and evaluation of robotic systems, particularly in tasks involving object manipulation across multiple environments.
+This dataset is released as part of the project:
+
+**PRTS: A Primitive Reasoning and Tasking System via Contrastive Representations**
+
+This dataset card describes the **PRTS Post-Training Data**, a high-quality robotic manipulation dataset designed for learning **primitive-level skills** and **long-horizon task execution**. The dataset covers both **dual-arm** and **single-arm** manipulation, including **fine manipulation** and **long-range operations**, and is curated in the **standard LeRobot format**.
+
+It serves as a **high-quality benchmark** for robotic manipulation, supporting research in policy learning, task decomposition, and embodied intelligence.
+
+---
 
 ## Dataset Details
 
 ### Dataset Description
 
-This dataset consists of robotic task data in the **LeRobot format**, covering various types of robotic manipulations such as **dual-arm**, **single-arm**, **long-range**, and **fine manipulation tasks**. It contains approximately 50GB of high-quality data collected from real-world robotic systems performing tasks in **household**, **office**, and **industrial** environments. The data includes detailed task frames including object handling, task coordination between two robotic arms, and complex manipulation actions.
+This dataset consists of robotic task data in the **LeRobot format**, covering diverse manipulation scenarios including:
+
+- **Dual-arm coordination**
+- **Single-arm manipulation**
+- **Long-horizon tasks**
+- **Fine-grained manipulation**
 
-- **Curated by:** TeleEmbodied AI Team
-- **License:** MIT License
+The dataset contains approximately **50GB** of high-quality data collected from real-world robotic systems across **household**, **office**, and **industrial** environments.
 
+Each trajectory captures detailed low-level control signals, including poses, joint states, and gripper actions, enabling learning of **primitive skills**, **action representations**, and **task-level policies**.
+
+- **Curated by:** TeleEmbodied AI Team  
+- **License:** MIT License  
+
+---
 
 ## Uses
 
 ### Direct Use
 
-This dataset is intended for the following use cases:
-- **Dual-arm** and **single-arm** robotic manipulation tasks.
-- **Object pick-and-place** and **fine manipulation** tasks.
-- **Long-range robotic operations** and task planning.
-- **Evaluating robotic coordination**, object handling, and task execution in real-world environments.
+This dataset is intended for:
+
+- Learning **dual-arm** and **single-arm** manipulation policies  
+- Training **primitive-level action representations**  
+- **Long-horizon task learning** and decomposition  
+- **Fine manipulation** skill learning  
+- Benchmarking **robotic manipulation models** in real-world settings  
+
+---
 
 ## Dataset Structure
 
-The dataset is curated in the **LeRobot format**. It contains data for multiple robotic tasks, each with the following key fields:
+The dataset follows the **standard LeRobot format**, widely used for robotic learning and policy training.
+
+Each task consists of structured trajectories with frame-wise robot states:
 
-- **`task_name`**: The name of the task (e.g., "Pick and Place Object").
-- **`task_id`**: A unique identifier for each task.
-- **`task_description`**: A brief description of the task.
-- **`frames`**: A collection of frames associated with the task. Each frame includes:
-  - **`t`**: Timestamp for the frame.
-  - **`pose`**: 6D pose of the robotic arms (for both left and right arms in dual-arm tasks).
-  - **`joints`**: Joint angles for the robotic arms (for both left and right arms in dual-arm tasks).
-  - **`gripper`**: Gripper state (open/close).
+- **`task_name`**: Task name (e.g., "Pick and Place Object")  
+- **`task_id`**: Unique task identifier  
+- **`task_description`**: Description of the task  
+- **`frames`**: Sequence of frames, each containing:
+  - **`t`**: Timestamp  
+  - **`pose`**: 6D end-effector pose (dual-arm or single-arm)  
+  - **`joints`**: Joint angles  
+  - **`gripper`**: Gripper state (open/close)  
 
-Here’s an example of what the data structure looks like:
+Example:
 
 ```json
 {