Baichuan-Omni-1.5

Hey everyone! Baichuan-Omni-1.5, a new open-source, omni-modal model from Baichuan AI, is now available. Key Features: 🌐 Multi-Modal: Processes text, image, video, and audio inputs; generates text and audio. 🏆 Strong Performance: Outperforms GPT-4o mini on multiple benchmarks, particularly in visual and audio tasks. ⚕️ Medical Capabilities: Shows significant promise in medical image understanding. 🔊 Advanced Audio: End-to-end audio processing, including high-quality speech synthesis (TTS) and automatic speech recognition (ASR). ✅ Open Source: Both base and fine-tuned models are available under a permissive license, allowing commercial use. 📊 Two New Evaluation Benchmarks: Baichuan also open-sourced two new evaluation benchmarks, OpenMM-Medical and OpenAudioBench. Baichuan-Omni-1.5 offers a powerful, open-source alternative for multi-modal AI development. While the fine-tuned model demonstrates exceptional strength in medical applications, the versatile base model provides a solid foundation for building a wide range of general-purpose applications.