Model description

Hi there,

Alibaba released the Ovis2 series a few months ago. They are a mix of Qwen2VL + Apple's AIMv2 VIT.

While their transformers-compatible code works, it would great for it to be upstreamed as to avoid trust_remote_code and to be adapted to transformer's processors for use in VLLM.

Open source status

• The model implementation is available
• The model weights are available

Provide useful links for the implementation

Transformers-compatible implementation here: https://huggingface.co/AIDC-AI/Ovis2-2B/tree/main