Image Understanding → Text Generation
Model: SmolVLM2-500M-Video-Instruct
Educational prototype inspired by multimodal LLM architectures. This is not the original NExT-GPT checkpoint.