Does anyone know of any Vision Language Models similar to DeepMind's Flamingo model that will take in a video clip a user provides the model along with a user-defined question in a prompt about what is happening in the video, and will generate high-quality answers to user's questions? I'm looking for models that are currently (or will soon be) available for commercial purposes. Ideally, the model should be able to generalize well to a variety of different vision-language tasks, or else at least perform well on tasks like event description, event understanding question answering, temporal/causal question answering, and visual dialogue.
Flamingo model paper for reference: https://arxiv.org/abs/2204.14198
