VisualAgentBench

VisualAgentBench copied to clipboard

**Describe the bug** Looking forward to the QuickStart!

zhangxi1997

请问有VAB-Mobile和WebArena-Lite这两个数据集生成的代码吗

1

wangjinghan666

When will the training datasets be released?

9

Thanks for this great work! When will the datasets be released, especially the training trajectory? Could you release the datasets in very near future?

zzjchen

[Bug/Assistance] Quickstart?

4

Hi, this looks great! Would love to get started using it -- any timeline for instructions on that?

sachit-menon

Thanks for releasing this great benchmark dataset!

1

Thank you for releasing such great work! We greatly appreciate your contribution to this domain and are very interested in the environments and datasets you've provided. Do you have any...

pokerme7777

[Feature] 会发布不需要容器就可以运行的环境代码吗？

1

类似于AgentBoard（https://github.com/hkust-nlp/AgentBoard ）。他们的环境可以直接运行而不需要容器。这可以方便非root用户使用visualagentbench。 我考虑agentbench不发布非docker版本的数据集是因为os任务是真的可能会影响运行环境的，但这几个任务貌似不涉及这个问题？

Fu-Dayuan

请问如何运行你们的评测代码呢？

2

需要安装docker环境吗？

LibertFan

overide -> override

eltociear

I want to write my own client using requests/http to access the server of vab-omnigibson (I use 'python -m src.start_task -a -s omnigibson 2' to start the server), while the...

lanseyege

我看了webRL和AutoGLM的Paper，autoGLM里提到使用webRL的训练框架，我猜它的数据集应该不是VAB开源的这部分数据集，极可能是涉及更多网站、特别是中文网站的任务吧？这部分数据有开源吗？

echoht