Multiple Videos as Input for NIM Prompt

AI & Data Science Computer Vision & Image Processing
,
1

When trying to pass multiple video uri’s through cosmos3 reasoner NIM I get the error:

“ERROR 07-29 09:28:54 [serving_chat.py:335] ValueError: At most 1 video(s) may be provided in one prompt.”

Full setup:

 {
                "role": "user",
                "content": [
                    {
                        "type": "video_url",
                        "video_url": {"url": left_mp4}
                    },

                    {
                        "type": "video_url",
                        "video_url": {"url": right_mp4}
                    },

                    {
                        "type": "text",
                        "text": self._caption_prompt
                    }
                ]
            }

There is no documentation that I could find on utilizing multiple video feeds in the reasoner, while for generator and predict there seems to be a native multi-camera schema to pass to the model, am I just passing multiple feeds in the wrong way or is this simply not possible to do without manipulating videos together in some way?