mlsanityData to 7 Oct 2026

SkyReels-V2: Infinite-length Film Generative Model

arXiv 2504.13074 ↗First posted 17 Apr 2025cs.CVcited by 218

On the map: Multimodal LLMs / vision-language models → Video understanding with MLLMs

also in 1 more
Abstract (arXiv)

Recent advances in video generation have been driven by diffusion models and autoregressive frameworks, yet critical challenges persist in harmonizing prompt adherence, visual quality, motion dynamics, and duration: compromises in motion dynamics to enhance temporal visual quality, constrained video duration (5-10 seconds) to prioritize resolution, and inadequate shot-aware generation stemming from general-purpose MLLMs' inability to interpret cinematic grammar, such as shot composition, actor expressions, and camera motions. These intertwined limitations hinder realistic long-form synthesis and professional film-style generation. To address these limitations, we propose SkyReels-V2, an Infinite-length Film Generative Model, that synergizes Multi-modal Large Language Model (MLLM), Multi-stage Pretraining, Reinforcement Learning, and Diffusion Forcing Framework. Firstly, we design a comprehensive structural representation of video that combines the general descriptions by the Multi-modal LLM and the detailed shot language by sub-expert models. Aided with human annotation, we then train a unified Video Captioner, named SkyCaptioner-V1, to efficiently label the video data. Secondly, we establish progressive-resolution pretraining for the fundamental video generation, followed by a four-stage post-training enhancement: Initial concept-balanced Supervised Fine-Tuning (SFT) improves baseline quality; Motion-specific Reinforcement Learning (RL) training with human-annotated and synthetic distortion data addresses dynamic artifacts; Our diffusion forcing framework with non-decreasing noise schedules enables long-video synthesis in an efficient search space; Final high-quality SFT refines visual fidelity. All the code and models are available at https://github.com/SkyworkAI/SkyReels-V2.

Key facts. Cited by 218 papers in the map (data through 7 Oct 2026). Placed in Video understanding with MLLMs, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 45 of its 58 references that are in the map. First posted on arXiv 17 Apr 2025.

Builds on

45 of 58 references.

References · 45

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  4. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  5. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  6. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,146
  7. 2022Scalable Diffusion Models with Transformerscited by 5,931
  8. 2022Flow Matching for Generative Modelingcited by 5,426
  9. 2025Qwen2.5-VL Technical Reportcited by 5,418
  10. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,094
  11. 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,229
  12. 2023Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasetscited by 2,526
33 more
  1. 2022Video Diffusion Modelscited by 2,280
  2. 2024CogVideoX: Text-to-Video Diffusion Models with An Expert Transformercited by 2,198
  3. 2023VBench: Comprehensive Benchmark Suite for Video Generative Modelscited by 1,755
  4. 2022Make-A-Video: Text-to-Video Generation without Text-Video Datacited by 1,725
  5. 2023AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuningcited by 1,418
  6. 2024MiniCPM-V: A GPT-4V Level MLLM on Your Phonecited by 1,061
  7. 2021MUSIQ: Multi-scale Image Quality Transformercited by 1,058
  8. 2023One-step Diffusion with Distribution Matching Distillationcited by 978
  9. 2025Cosmos World Foundation Model Platform for Physical AIcited by 833
  10. 2023Diffusion Model Alignment Using Direct Preference Optimizationcited by 767
  11. 2019Advantage-Weighted Regression: Simple and Scalable Off-Policy Reinforcement Learningcited by 750
  12. 2024Improved Distribution Matching Distillation for Fast Image Synthesiscited by 648
  13. 2024Open-Sora: Democratizing Efficient Video Production for Allcited by 637
  14. 2024Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusioncited by 634
  15. 2022Exploring Video Quality Assessment on User Generated Contents from Aesthetic and Technical Perspectivescited by 314
  16. 2019Character Region Awareness for Text Detectioncited by 273
  17. 2025Improving Video Generation with Human Feedbackcited by 267
  18. 2023Using Human Feedback to Fine-tune Diffusion Models without Any Reward Modelcited by 246
  19. 2024Timestep Embedding Tells: It's Time to Cache for Video Diffusion Modelcited by 245
  20. 2023DeepSpeed Ulysses: System Optimizations for Enabling Training of Extreme Long Sequence Transformer Modelscited by 239
  21. 2022FAST-VQA: Efficient End-to-end Video Quality Assessment with Fragment Samplingcited by 179
  22. 2022A Self-Supervised Descriptor for Image Copy Detectioncited by 175
  23. 2020TransNet V2: An effective deep network architecture for fast shot transition detectioncited by 166
  24. 2025Open-Sora 2.0: Training a Commercial-Level Video Generation Model in 200kcited by 157
  25. 2023UniMax: Fairer and more Effective Language Sampling for Large-Scale Multilingual Pretrainingcited by 147
  26. 2023ARNIQA: Learning Distortion Manifold for Image Quality Assessmentcited by 87
  27. 2025SkyReels-A2: Compose Anything in Video Diffusion Transformerscited by 72
  28. 2024CV-VAE: A Compatible Video VAE for Latent Generative Video Modelscited by 68
  29. 2024HumanVid: Demystifying Training Data for Camera-controllable Human Image Animationcited by 63
  30. 2024WF-VAE: Enhancing Video VAE by Wavelet-Driven Energy Flow for Latent Video Diffusion Modelcited by 55
  31. 2025AR-Diffusion: Asynchronous Video Generation with Auto-Regressive Diffusioncited by 53
  32. 2023From Static to Dynamic: Adapting Landmark-Aware Image Models for Facial Expression Recognition in Videoscited by 39
  33. 2025SkyReels-A1: Expressive Portrait Animation in Video Diffusion Transformerscited by 33

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 74; 2026: 144.

218 papers.

Sort:

By area:

  1. 2025Flow-GRPO: Training Flow Matching Models via Online RLcited by 605
  2. 2025Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusioncited by 569
  3. 2025LongLive: Real-time Interactive Long Video Generationcited by 237
  4. 2025Rolling Forcing: Autoregressive Long Video Diffusion in Real Timecited by 209
  5. 2025Self-Forcing++: Towards Minute-Scale High-Quality Video Generationcited by 176
  6. 2025Matrix-game 2.0: An open-source real-time and streaming interactive world modelcited by 174
  7. 2026Causal Forcing: Autoregressive Diffusion Distillation Done Right for High-Quality Real-Time Interactive Video Generationcited by 159
  8. 2026Advancing Open-source World Modelscited by 144
  9. 2025Video World Models with Long-term Spatial Memorycited by 124
  10. 2025SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformercited by 91
  11. 2025Mixture of Contexts for Long Video Generationcited by 90
  12. 2025RELIC: Interactive Video World Model with Long-Horizon Memorycited by 89
38 more
  1. 2025Autoregressive Adversarial Post-Training for Real-Time Interactive Video Generationcited by 84
  2. 2025Infinity-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rolloutcited by 78
  3. 2025Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillationcited by 75
  4. 2025Yume: An Interactive World Generation Modelcited by 73
  5. 2025Towards Understanding Camera Motions in Any Videocited by 62
  6. 2026Helios: Real Real-Time Long Video Generation Modelcited by 56
  7. 2026SANA-WM: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion Transformercited by 50
  8. 2026Context Forcing: Consistent Autoregressive Video Generation with Long Contextcited by 47
  9. 2025End-to-End Training for Autoregressive Video Diffusion via Self-Resamplingcited by 45
  10. 2025FlashVSR: Towards Real-Time Diffusion-Based Streaming Video Super-Resolutioncited by 40
  11. 2025MemFlow: Flowing Adaptive Memory for Consistent and Efficient Long Video Narrativescited by 40
  12. 2025VideoSSM: Autoregressive Long Video Generation with Hybrid State-Space Memorycited by 39
  13. 2026Infinite Worlds with Versatile Interactionscited by 37
  14. 2025Spatia: Video Generation with Updatable Spatial Memorycited by 33
  15. 2025DenseDPO: Fine-Grained Temporal Preference Optimization for Video Diffusion Modelscited by 32
  16. 2026LoL: Longer than Longer, Scaling Video Generation to Hourcited by 29
  17. 2026Rolling Sink: Bridging Limited-Horizon Training and Open-Ended Testing in Autoregressive Video Diffusioncited by 28
  18. 2025Stand-In: A Lightweight and Plug-and-Play Identity Control for Video Generationcited by 27
  19. 2025StreamDiffusionV2: A Streaming System for Dynamic and Interactive Video Generationcited by 27
  20. 2026LongLive-2.0: An NVFP4 Parallel Infrastructure for Long Video Generationcited by 25
  21. 2025Skyra: AI-Generated Video Detection via Grounded Artifact Reasoningcited by 24
  22. 2026AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map Distillationcited by 23
  23. 2026Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Modelscited by 23
  24. 2026VideoMemory: Toward Consistent Video Generation via Memory Integrationcited by 22
  25. 2025NarrLV: Towards a Comprehensive Narrative-Centric Evaluation for Long Video Generationcited by 21
  26. 2025Hierarchical Fine-grained Preference Optimization for Physically Plausible Video Generationcited by 20
  27. 2026MemRoPE: Training-Free Infinite Video Generation via Evolving Memory Tokenscited by 20
  28. 2025LoViC: Efficient Long Video Generation with Context Compressioncited by 19
  29. 2025Macro-from-Micro Planning for High-Quality and Parallelized Autoregressive Long Video Generationcited by 19
  30. 2026SkyReels-V3 Technique Reportcited by 19
  31. 2026Flow caching for autoregressive video generationcited by 19
  32. 2026DreamID-Omni: Unified Framework for Controllable Human-Centric Audio-Video Generationcited by 19
  33. 2025X-Streamer: Unified Human World Modeling with Audiovisual Interactioncited by 16
  34. 2025Cut2Next: Generating Next Shot via In-Context Tuningcited by 15
  35. 2025DC-VideoGen: Efficient Video Generation with Deep Compression Video Autoencodercited by 15
  36. 2025Towards One-step Causal Video Generation via Adversarial Self-Distillationcited by 15
  37. 2025Simulating the Visual World with Artificial Intelligence: A Roadmapcited by 15
  38. 2026Efficient Autoregressive Video Diffusion with Dummy Headcited by 15

50 most cited of 218 All 218 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Captain Cinema: Towards Short Movie Generation
  2. 2026Auteur: Language-Driven Cinematographic Framing for Human-Centric Video Generation
  3. 2024Video Is Worth a Thousand Images: Exploring the Latest Trends in Long Video Generation
  4. 2025Dynamic-I2V: Exploring Image-to-Video Generation Models via Multimodal LLM
  5. 2026CineWeaver: Training-Free Reference-Controllable Multi-Shot Long Video Generation for Cinematic Storytelling
  6. 2025Anchored Video Generation: Decoupling Scene Construction and Temporal Synthesis in Text-to-Video Diffusion Models
  7. 2023Gen-L-Video: Multi-Text to Long Video Generation via Temporal Co-Denoising
  8. 2024Can video generation replace cinematographers? Research on the cinematic language of generated video