Open Issues Need Help
View All on GitHubAI Summary: Develop a module to automatically create datasets from `rai_benchmark` traces for fine-tuning smaller language/vision-language models (LLMs/VLMs). The module should be independent of `rai_benchmark` to allow future use with other deployments. The dataset should include relevant metadata for flexible fine-tuning with various models, though prompt structure adjustments might be handled in a separate pre-processing step.
RAI is a vendor-agnostic agentic framework for robotics, utilizing ROS 2 tools to perform complex actions, defined scenarios, free interface execution, log summaries, voice interaction and more.