by idea-research
Empower LLMs with fine-grained visual understanding โ detect, localize, and describe anything in images with natural language prompts.