Ovis1.6 Llama3.2 3B It is recommended to use the latest version: Ovis2. Introduction GitHub Demo Paper We are thrilled to announce the open sourcing of Ovis1.6 Llama3.2 3B , an integral part of the Ovis1.6 family. This cutting edge model currently sets the benchmark as the state of the art (SOTA) solution for edge side multimodal tasks. The Ovis family employs an innovative Multimodal Large Language Model (MLLM) architecture, specifically designed to structurally align visual and textual embeddings. Ovis1.6 Llama3.2 3B excels in common industry benchmarks, surpassing numerous open source and proprietary multimodal models. Moreover, it is also particularly well suited for local intelligence, on device computing, and edge computing scenarios. Model Built upon Ovis1.5, Ovis1.6 further enhances high resolution image processing, is trained on a larger, more diverse, and higher quality dataset, and refines the training process with DPO training following instruction tuning. Ovis MLLMs ViT LLM Model Weights Demo : : : : : : : : : Ovis1.6 Gemma2 27B Siglip 400M Gemma2 27B It Huggingface Ovis1.6 Gemma2 9B Siglip 400M Gemma2 9B It Huggingface Space Ovis1.6 Llama3.2 3B Siglip 400M Llama 3.2 3…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy