CineMobile: On-Device Image-to-Video Diffusion for Cinematic Camera Motion Generation

Hugging Face Daily Papers Papers

Summary

This paper presents CineMobile, a method for efficient on-device image-to-video generation that achieves a 40x speedup over the teacher model through distillation-guided pruning, diffusion distillation, and hybrid quantization, enabling cinematic camera motion effects on mobile devices.

The growing demand for image-to-video creation on mobile devices has increasingly focused on cinematic motion effects like bullet time, dolly zoom, slow motion, etc. While Diffusion Transformers (DiTs) exhibit strong performance in video generation, their large parameter sizes and multi-step iterative denoising processes lead to substantial computational overhead, making efficient generation on mobile devices challenging. We propose CineMobile to bridge the gap. In particular, CineMobile adopts a three-fold optimization strategy: (1) leveraging a distillation-guided pruning approach to derive a compact yet efficient model that retains the essential video generation capabilities required for cinematic effects; (2) optimizing the compressed model into a 4-step generator via a combination of diffusion distillation and reinforcement learning; (3) employing a hybrid post-training quantization strategy to compress the model footprint to under 1 GB. Experimental results show that compared to the teacher model with the Wan 2.1 architecture, CineMobile achieves a 40x speedup in generation while maintaining comparable visual quality. Specifically, CineMobile generates 49-frame 480p videos with a per-step denoising latency of 0.6s on an NVIDIA H200 GPU and 20s on the MediaTek Dimensity 8400 Ultimate 5G platform, with a peak memory usage of 1.8 GB, demonstrating its practical applicability for mobile-based image-to-video creation.
Original Article
View Cached Full Text

Cached at: 07/10/26, 06:16 AM

Paper page - CineMobile: On-Device Image-to-Video Diffusion for Cinematic Camera Motion Generation

Source: https://huggingface.co/papers/2607.03803

Abstract

CineMobile enables efficient image-to-video generation on mobile devices through distillation-guided pruning, diffusion distillation, and hybrid quantization techniques while maintaining visual quality and achieving significant speedup.

The growing demand for image-to-video creation on mobile devices has increasingly focused oncinematic motion effectslike bullet time, dolly zoom, slow motion, etc. WhileDiffusion Transformers(DiTs) exhibit strong performance invideo generation, their large parameter sizes and multi-step iterative denoising processes lead to substantial computational overhead, making efficient generation on mobile devices challenging. We propose CineMobile to bridge the gap. In particular, CineMobile adopts a three-fold optimization strategy: (1) leveraging adistillation-guided pruningapproach to derive a compact yet efficient model that retains the essentialvideo generationcapabilities required for cinematic effects; (2) optimizing the compressed model into a 4-step generator via a combination ofdiffusion distillationandreinforcement learning; (3) employing ahybrid post-training quantizationstrategy to compress the model footprint to under 1 GB. Experimental results show that compared to the teacher model with the Wan 2.1 architecture, CineMobile achieves a 40x speedup in generation while maintaining comparable visual quality. Specifically, CineMobile generates 49-frame 480p videos with a per-step denoising latency of 0.6s on an NVIDIA H200 GPU and 20s on the MediaTek Dimensity 8400 Ultimate 5G platform, with a peak memory usage of 1.8 GB, demonstrating its practical applicability for mobile-based image-to-video creation.

View arXiv pageView PDFAdd to collection

Get this paper in your agent:

hf papers read 2607\.03803

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2607.03803 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.03803 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.03803 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

MobileMoE: Scaling On-Device Mixture of Experts

Hugging Face Daily Papers

MobileMoE introduces efficient on-device mixture-of-experts language models with sub-billion parameters, achieving better performance and efficiency than dense baselines and existing MoE models. The models are trained on open-source datasets and demonstrate significant speedups on commodity smartphones.