MiniCPM-o 4.5: Watching, Listening and Proactive Speaking Simultaneously

💫 Ready to build with MiniCPM-o 4.5? We’ve made it #developer-ready with full support for llama.cpp, ollama, vLLM, SGLang, LLaMA-Factory, etc. 💻 We also open source a new high-performing llama.cpp-omni engine together with an interactive Demo, to bring full-duplex omni experience directly to your local devices like MacBook! 🚀 Try it on @Hugging Face now 👉🏻 huggingface.co/openbmb/MiniCPM-o-4_5