xavierpuigf/virtualhome

• highreadme#1

  Reposition README opening to emphasize synthetic data generation

  原因：

  当前

  VirtualHome is an interactive platform to simulate complex household activities via programs. Key aspect of VirtualHome is that it allows complex interactions with the environment, such as picking up objects, switching on/off appliances, opening appliances, etc. Our simulator can easily be called with a Python API: write the activity as a simple sequence of instructions which then get rendered in VirtualHome. You can choose between different agents and environments, as well as modify environments on the fly. You can also stream different ground-truth such as time-stamped actions, instance/semantic segmentation, and optical flow and depth. The platform allows to simulate multi-agent activities and can serve as an environment to train agents for embodied AI tasks.

  复制粘贴的修复

  复制

  **VirtualHome** is an interactive platform to simulate complex household activities via programs, serving as a robust environment for embodied AI research and **synthetic data generation**. It provides rich ground-truth data, including instance/semantic segmentation, optical flow, and depth, alongside its core capability for complex interactions with the environment (picking up objects, switching appliances). Our simulator, easily called with a Python API, allows writing activities as simple instruction sequences, rendered in VirtualHome, and supports multi-agent activities and dynamic environment modification.

  待办完成忽略
• mediumtopics#2

  Add specific topics for synthetic data and embodied AI

  原因：

  当前

  computer-vision, deep-learning, graph, multi-agent, reinforcement-learning, simulator, unity

  复制粘贴的修复

  复制

  computer-vision, deep-learning, graph, multi-agent, reinforcement-learning, simulator, unity, synthetic-data, data-generation, segmentation, depth-estimation, embodied-ai

  待办完成忽略
• lowreadme#3

  Add a 'Key Capabilities' section to the README

  原因：

  复制粘贴的修复

  复制

  ## Key Capabilities
  
  *   **Programmatic Control:** Define complex household activities via a simple Python API.
  *   **Multi-Agent Simulation:** Support for multiple AI agents interacting within the environment.
  *   **Rich Ground-Truth Data:** Stream instance/semantic segmentation, optical flow, and depth for synthetic data generation.
  *   **Interactive Environments:** Agents can pick up objects, switch appliances, open/close containers, and modify environments dynamically.
  *   **Procedural Generation:** Create infinite unique environments for diverse training scenarios.
  *   **Embodied AI Training:** An ideal environment for training agents in complex embodied AI tasks.

  待办完成忽略

本次扫描解析到的品类 GEO 通道：google/gemini-2.5-flash, deepseek/deepseek-v4-flash

• 品类问题

  What simulation environments support training multiple AI agents in interactive household settings?

  你：第 3 位
  AI 推荐顺序：

  1. AI2-THOR
  2. Habitat
  3. VirtualHome ← 你
  4. iGibson
  5. MINERVA
  查看 AI 完整回答
• 品类问题

  Looking for a simulator to generate synthetic data for embodied AI, including segmentation and depth.

  你：未被推荐
  AI 推荐顺序：

  1. NVIDIA Isaac Sim
  2. Unity 3D
  3. Perception Package
  4. Unreal Engine
  5. AirSim
  6. CARLA
  7. Habitat-Sim
  8. Gazebo

  AI 推荐了 8 个替代方案，却始终没点名 xavierpuigf/virtualhome。这就是要补上的差距。

  查看 AI 完整回答

• Metadata completeness

  pass

• README presence

  pass

• Compared to common alternatives in this category, what is the core differentiator of xavierpuigf/virtualhome?

  pass
  ✓AI 明确点名了 xavierpuigf/virtualhome

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• If a team adopts xavierpuigf/virtualhome in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI 明确点名了 xavierpuigf/virtualhome

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• In one sentence, what problem does the repo xavierpuigf/virtualhome solve, and who is the primary audience?

  pass
  ✓AI 明确点名了 xavierpuigf/virtualhome

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？