Scaling RL to Long Videos

• URL: http://arxiv.org/abs/2507.07966v3
• Date: Wed, 30 Jul 2025 16:55:33 GMT
• Title: Scaling RL to Long Videos
• Authors: Yukang Chen, Wei Huang, Baifeng Shi, Qinghao Hu, Hanrong Ye, Ligeng Zhu, Zhijian Liu, Pavlo Molchanov, Jan Kautz, Xiaojuan Qi, Sifei Liu, Hongxu Yin, Yao Lu, Song Han,
• Abstract summary: LongVILA-R1-7B achieves strong performance on video benchmarks, reaching 65.1% and 71.1% accuracy on VideoMME without and with subtitles, respectively.<n>LongVILA-R1-7B supports processing up to 8,192 video frames per video, and FPS settings.<n>We release our training system for public availability that supports RL training on various modalities.
• Score: 107.41198639507255
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: We introduce a full-stack framework that scales up reasoning in vision-language models (VLMs) to long videos, leveraging reinforcement learning. We address the unique challenges of long video reasoning by integrating three critical components: (1) a large-scale dataset, LongVideo-Reason, comprising 104K long video QA pairs with high-quality reasoning annotations across diverse domains such as sports, games, and vlogs; (2) a two-stage training pipeline that extends VLMs with chain-of-thought supervised fine-tuning (CoT-SFT) and reinforcement learning (RL); and (3) a training infrastructure for long video RL, named Multi-modal Reinforcement Sequence Parallelism (MR-SP), which incorporates sequence parallelism and a vLLM-based engine tailored for long video, using cached video embeddings for efficient rollout and prefilling. In our experiments, LongVILA-R1-7B achieves strong performance on video benchmarks, reaching 65.1% and 71.1% accuracy on VideoMME without and with subtitles, respectively, and consistently outperforming LongVILA-7B across multiple benchmarks. Moreover, LongVILA-R1-7B supports processing up to 8,192 video frames per video, and configurable FPS settings. Notably, our MR-SP system achieves up to 2.1x speedup on long video RL training. In addition, we release our training system for public availability that supports RL training on various modalities (video, text, and audio), various models (VILA and Qwen series), and even image and video generation models. On a single A100 node (8 GPUs), it supports RL training on hour-long videos (e.g., 3,600 frames).

Related papers

• ARC-Hunyuan-Video-7B: Structured Video Comprehension of Real-World Shorts [56.75723197779384]
  ARC-Hunyuan-Video is a multimodal model that processes visual, audio, and textual signals end-to-end for structured comprehension.<n>Our model is capable of multi-granularity timestamped video captioning and summarization, open-ended video question answering, temporal video grounding, and video reasoning.
  arXiv  Detail & Related papers  (2025-07-28T15:52:36Z)
• Unleashing Hour-Scale Video Training for Long Video-Language Understanding [61.717205915329664]
  We present VideoMarathon, a large-scale hour-long video instruction-following dataset.<n>This dataset includes around 9,700 hours of long videos sourced from diverse domains, ranging from 3 to 60 minutes per video.<n>We propose Hour-LLaVA, a powerful and efficient Video-LMM for hour-scale video-language modeling.
  arXiv  Detail & Related papers  (2025-06-05T17:59:04Z)
• ViaRL: Adaptive Temporal Grounding via Visual Iterated Amplification Reinforcement Learning [68.76048244253582]
  We introduce ViaRL, the first framework to leverage rule-based reinforcement learning (RL) for optimizing frame selection in video understanding.<n>ViaRL utilizes the answer accuracy of a downstream model as a reward signal to train a frame selector through trial-and-error.<n>ViaRL consistently delivers superior temporal grounding performance and robust generalization across diverse video understanding tasks.
  arXiv  Detail & Related papers  (2025-05-21T12:29:40Z)
• Time-R1: Post-Training Large Vision Language Model for Temporal Video Grounding [57.26400319795876]
  Temporal Video Grounding (TVG) is a core challenge in long-form video understanding.<n>Recent Large Vision-Language Models (LVLMs) have shown early promise in tackling TVG through supervised fine-tuning.<n>We propose a novel post-training framework that enhances the generalization capabilities of LVLMs via reinforcement learning.
  arXiv  Detail & Related papers  (2025-03-17T17:04:20Z)
• HLV-1K: A Large-scale Hour-Long Video Benchmark for Time-Specific Long Video Understanding [20.184894298462652]
  We build a large-scale hour-long long video benchmark, HLV-1K, designed to evaluate long video understanding models.<n>HLV-1K comprises 1009 hour-long videos with 14,847 high-quality question answering (QA) and multi-choice question asnwering (MCQA)<n>We evaluate our benchmark using existing state-of-the-art methods and demonstrate its value for testing deep long video understanding capabilities at different levels and for various tasks.
  arXiv  Detail & Related papers  (2025-01-03T05:32:37Z)
• VideoChat-Flash: Hierarchical Compression for Long-Context Video Modeling [43.485687038460895]
  Long-context video modeling is critical for multimodal large language models (MLLMs)<n>This paper aims to address this issue from aspects of model architecture, training data, training strategy and evaluation benchmark.<n>We build a powerful video MLLM named VideoChat-Flash, which shows a leading performance on both mainstream long and short video benchmarks.
  arXiv  Detail & Related papers  (2024-12-31T18:01:23Z)
• LongVILA: Scaling Long-Context Visual Language Models for Long Videos [86.28679075537089]
  LongVILA is a full-stack solution for long-context visual-language models.<n>LongVILA efficiently extends the number of video frames of VILA from 8 to 2048, achieving 99.8% accuracy in 6,000-frame (more than 1 million tokens) video needle-in-a-haystack.
  arXiv  Detail & Related papers  (2024-08-19T17:48:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.