Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.
Existing VLMs can be very useful for robotics but they struggle to answer multiple-choice questions about robot scenes. This paper explores this gap and provides benchmarks for improvement. Kudos to the team of authors led by postdocs Shuangyu Xie and Eric Chen! https://twitter.com/shuangyu0_0xie/status/2076737025437528268
What makes a VLM good at real robotic tasks? We present a new Robot Question Answering (RQA) benchmark across domains relevant for robotics and study their failure modes. TLDR: Start with Gemini and Qwen (open), and benchmark on RQA! https://berkeleyautomation.github.io/robovista/ 🧵:
Existing vision-language models struggle with complex, real-world robot scenes
Existing VLMs can be very useful for robotics but they struggle to answer multiple-choice questions about robot scenes. This paper explores this gap and provides benchmarks for improvement. Kudos to the team of authors led by postdocs Shuangyu Xie and Eric Chen! https://twitter.com/shuangyu0_0xie/status/2076737025437528268
What makes a VLM good at real robotic tasks? We present a new Robot Question Answering (RQA) benchmark across domains relevant for robotics and study their failure modes. TLDR: Start with Gemini and Qwen (open), and benchmark on RQA! https://berkeleyautomation.github.io/robovista/ 🧵:
Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.