martin-ev/abo_vgoq

---
license: cc-by-nc-4.0
source_datasets:
- ABO-dataset
---

This synthetic dataset is built on top of the [Amazon Berkeley Objects (ABO) dataset](https://amazon-berkeley-objects.s3.amazonaws.com/index.html), which contains Amazon products with product listings and catalog images.

The fields `object_question`, `most_relevant_image_path`, `grounding_mask`, and `grounding_evidential_quality` contain our new synthetically generated questions/selection/grounding/annotations.

**License (source dataset):** The source dataset (Amazon Berkeley Objects (ABO)) is licensed under the [Creative Commons Attribution 4.0 International License](https://creativecommons.org/licenses/by/4.0/).

**License (our synthetic questions/selection/grounding/annotations):** Our newly generated questions/selection/grounding/annotations are licensed under the [Creative Commons Attribution-NonCommercial 4.0 International License](https://creativecommons.org/licenses/by-nc/4.0/). Because the new questions/selection/grounding/annotations are generated using Claude 3.7 Sonnet (and other models), [Anthropic's usage policies](https://www.anthropic.com/legal/aup) also apply, see especially [model training with Claude's outputs](https://support.claude.com/en/articles/12326764-can-i-use-my-outputs-to-train-an-ai-model).

许可协议：CC BY-NC 4.0
源数据集：
- ABO数据集（ABO-dataset）

本合成数据集基于[亚马逊伯克利物体（Amazon Berkeley Objects，ABO）数据集](https://amazon-berkeley-objects.s3.amazonaws.com/index.html)构建，该数据集收录了带有商品详情页与目录图片的亚马逊在售商品。

字段`object_question`（对象问题）、`most_relevant_image_path`（最相关图片路径）、`grounding_mask`（锚定掩码）以及`grounding_evidential_quality`（锚定证据质量），包含我们全新合成生成的问题、选样、锚定任务及标注数据。

**源数据集许可说明**：源数据集（亚马逊伯克利物体（ABO）数据集）采用[知识共享署名4.0国际许可协议](https://creativecommons.org/licenses/by/4.0/)进行授权。

**合成标注许可说明**：我们新生成的问题、选样、锚定任务及标注数据采用[知识共享署名-非商业性使用4.0国际许可协议](https://creativecommons.org/licenses/by-nc/4.0/)进行授权。由于该类新生成内容通过Claude 3.7 Sonnet（及其他模型）生成，[Anthropic的使用政策](https://www.anthropic.com/legal/aup)同样适用，尤其请参阅[使用Claude输出训练AI模型](https://support.claude.com/en/articles/12326764-can-i-use-my-outputs-to-train-an-ai-model)。