This PR significantly improves the dataset card for ByteCameraDepth by:
- Adding `task_categories: - depth-estimation` to the metadata, making the dataset discoverable by its primary AI task.
- Including relevant `tags: - robotics`, `- 3d`, and `- simulation` to further enhance searchability, reflecting the dataset's context in robotic manipulation and 3D geometry perception, often used for bridging the sim-to-real gap.
- Adding prominent links to the associated paper ([Manipulation as in Simulation: Enabling Accurate Geometry Perception in Robots](https://huggingface.co/papers/2509.02530)), the project page ([https://manipulation-as-in-simulation.github.io/](https://manipulation-as-in-simulation.github.io/)), and the main GitHub repository ([https://github.com/ByteDance-Seed/manip-as-in-sim-suite](https://github.com/ByteDance-Seed/manip-as-in-sim-suite)) for easy access to related resources.
- Incorporating a "Sample Usage" section with a code snippet from the GitHub README, demonstra…
