VideoLLM-online: Online Video Large Language Model for Streaming Video
This addresses the inefficiency of existing models for real-time video processing, which is crucial for applications like live streaming and surveillance, though it is incremental as it builds upon existing LLMs like Llama-2/Llama-3.
The paper tackles the problem of large multimodal models being inefficient for streaming video inputs by proposing the Learning-In-Video-Stream (LIVE) framework, which enables real-time conversation in continuous video streams, achieving over 10 FPS on a 5-minute clip with an A100 GPU and state-of-the-art performance on offline benchmarks.
Recent Large Language Models have been enhanced with vision capabilities, enabling them to comprehend images, videos, and interleaved vision-language content. However, the learning methods of these large multimodal models typically treat videos as predetermined clips, making them less effective and efficient at handling streaming video inputs. In this paper, we propose a novel Learning-In-Video-Stream (LIVE) framework, which enables temporally aligned, long-context, and real-time conversation within a continuous video stream. Our LIVE framework comprises comprehensive approaches to achieve video streaming dialogue, encompassing: (1) a training objective designed to perform language modeling for continuous streaming inputs, (2) a data generation scheme that converts offline temporal annotations into a streaming dialogue format, and (3) an optimized inference pipeline to speed up the model responses in real-world video streams. With our LIVE framework, we built VideoLLM-online model upon Llama-2/Llama-3 and demonstrate its significant advantages in processing streaming videos. For instance, on average, our model can support streaming dialogue in a 5-minute video clip at over 10 FPS on an A100 GPU. Moreover, it also showcases state-of-the-art performance on public offline video benchmarks, such as recognition, captioning, and forecasting. The code, model, data, and demo have been made available at https://showlab.github.io/videollm-online.