GigaVideo-1: Advancing Video Generation via Automatic Feedback with 4 GPU-Hours Fine-Tuning

• URL: http://arxiv.org/abs/2506.10639v1
• Date: Thu, 12 Jun 2025 12:25:37 GMT
• Title: GigaVideo-1: Advancing Video Generation via Automatic Feedback with 4 GPU-Hours Fine-Tuning
• Authors: Xiaoyi Bao, Jindi Lv, Xiaofeng Wang, Zheng Zhu, Xinze Chen, YuKun Zhou, Jiancheng Lv, Xingang Wang, Guan Huang,
• Abstract summary: GigaVideo-1 is an efficient fine-tuning framework that advances video generation without additional human supervision.<n>We focus on two key aspects of the fine-tuning process: data and optimization.<n>Experiments show that GigaVideo-1 consistently improves performance on almost all the dimensions with an average gain of about 4% using only 4 GPU-hours.
• Score: 38.85475786611648
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent progress in diffusion models has greatly enhanced video generation quality, yet these models still require fine-tuning to improve specific dimensions like instance preservation, motion rationality, composition, and physical plausibility. Existing fine-tuning approaches often rely on human annotations and large-scale computational resources, limiting their practicality. In this work, we propose GigaVideo-1, an efficient fine-tuning framework that advances video generation without additional human supervision. Rather than injecting large volumes of high-quality data from external sources, GigaVideo-1 unlocks the latent potential of pre-trained video diffusion models through automatic feedback. Specifically, we focus on two key aspects of the fine-tuning process: data and optimization. To improve fine-tuning data, we design a prompt-driven data engine that constructs diverse, weakness-oriented training samples. On the optimization side, we introduce a reward-guided training strategy, which adaptively weights samples using feedback from pre-trained vision-language models with a realism constraint. We evaluate GigaVideo-1 on the VBench-2.0 benchmark using Wan2.1 as the baseline across 17 evaluation dimensions. Experiments show that GigaVideo-1 consistently improves performance on almost all the dimensions with an average gain of about 4% using only 4 GPU-hours. Requiring no manual annotations and minimal real data, GigaVideo-1 demonstrates both effectiveness and efficiency. Code, model, and data will be publicly available.

Related papers

• Seedance 1.0: Exploring the Boundaries of Video Generation Models [71.26796999246068]
  Seedance 1.0 is a high-performance and inference-efficient video foundation generation model.<n>It integrates multi-source curation data augmented with precision and meaningful video captioning.<n>Seedance 1.0 can generate a 5-second video at 1080p resolution only with 41.4 seconds ( NVIDIA-L20)
  arXiv  Detail & Related papers  (2025-06-10T17:56:11Z)
• Wan: Open and Advanced Large-Scale Video Generative Models [83.03603932233275]
  Wan is a suite of video foundation models designed to push the boundaries of video generation.<n>We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community.
  arXiv  Detail & Related papers  (2025-03-26T08:25:43Z)
• SlowFast-LLaVA-1.5: A Family of Token-Efficient Video Large Language Models for Long-Form Video Understanding [70.84791600974337]
  We introduce SlowFast-LLaVA-1.5 (abbreviated as SF-LLaVA-1.5), a family of video large language models (LLMs)<n>We incorporate the two-stream SlowFast mechanism into a streamlined training pipeline.<n>We perform joint video-image training on a carefully curated data mixture of only publicly available datasets.
  arXiv  Detail & Related papers  (2025-03-24T17:59:07Z)
• AMD-Hummingbird: Towards an Efficient Text-to-Video Model [12.09360569154206]
  Text-to-Video (T2V) generation has attracted significant attention for its ability to synthesize realistic videos from textual descriptions.<n>Most prior work prioritizes visual fidelity while overlooking the need for smaller, more efficient models suitable for real-world deployment.<n>We propose a lightweight T2V framework, termed Hummingbird, which prunes existing models and enhances visual quality through visual feedback learning.
  arXiv  Detail & Related papers  (2025-03-24T11:13:33Z)
• TinyLLaVA-Video: Towards Smaller LMMs for Video Understanding with Group Resampler [10.92767902813594]
  We introduce TinyLLaVA-Video, a lightweight yet powerful video understanding model with approximately 3.6B parameters.<n>The cornerstone of our design is the video-level group resampler, a novel mechanism that significantly reduces and controls the number of visual tokens at the video level.<n>TinyLLaVA-Video demonstrates exceptional efficiency, requiring only one day of training on 8 A100-40G GPUs.
  arXiv  Detail & Related papers  (2025-01-26T13:10:12Z)
• VideoSAVi: Self-Aligned Video Language Models without Human Supervision [0.6854849895338531]
  VideoSAVi is a self-training pipeline that enables Video-LLMs to reason over video content without external supervision.<n>VideoSAVi achieves state-of-the-art performance on MVBench (74.0%) and delivers significant improvements.<n>Our model-agnostic approach is computationally efficient, requiring only 32 frames.
  arXiv  Detail & Related papers  (2024-12-01T00:33:05Z)
• OpenHumanVid: A Large-Scale High-Quality Dataset for Enhancing Human-Centric Video Generation [27.516068877910254]
  We introduce OpenHumanVid, a large-scale and high-quality human-centric video dataset.<n>Our findings yield two critical insights: First, the incorporation of a large-scale, high-quality dataset substantially enhances evaluation metrics for generated human videos.<n>Second, the effective alignment of text with human appearance, human motion, and facial motion is essential for producing high-quality video outputs.
  arXiv  Detail & Related papers  (2024-11-28T07:01:06Z)
• VBench++: Comprehensive and Versatile Benchmark Suite for Video Generative Models [111.5892290894904]
  VBench is a benchmark suite that dissects "video generation quality" into specific, hierarchical, and disentangled dimensions. We provide a dataset of human preference annotations to validate our benchmarks' alignment with human perception. VBench++ supports evaluating text-to-video and image-to-video.
  arXiv  Detail & Related papers  (2024-11-20T17:54:41Z)
• NVDS+: Towards Efficient and Versatile Neural Stabilizer for Video Depth Estimation [58.21817572577012]
  Video depth estimation aims to infer temporally consistent depth. We introduce NVDS+ that stabilizes inconsistent depth estimated by various single-image models in a plug-and-play manner. We also elaborate a large-scale Video Depth in the Wild dataset, which contains 14,203 videos with over two million frames.
  arXiv  Detail & Related papers  (2023-07-17T17:57:01Z)

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.