Enregistré dans:
Détails bibliographiques
Auteurs principaux: Meng, Rang, Wu, Weipeng, Li, Yuming, Ma, Chenguang
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2602.13669
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914497007976448
author Meng, Rang
Wu, Weipeng
Li, Yuming
Ma, Chenguang
author_facet Meng, Rang
Wu, Weipeng
Li, Yuming
Ma, Chenguang
contents Recent multi-modal video generation models have achieved high visual quality, but their prohibitive latency and limited temporal stability hinder real-time deployment. Streaming inference exacerbates these issues, leading to pronounced multimodal degradation, such as spatial blurring, temporal drift, and lip desynchronization, which creates an unresolved efficiency-performance trade-off. To this end, we propose EchoTorrent, a novel schema with a fourfold design: (1) Multi-Teacher Training fine-tunes a pre-trained model on distinct preference domains to obtain specialized domain experts, which sequentially transfer domain-specific knowledge to a student model; (2) Adaptive CFG Calibration (ACC-DMD), which calibrates the audio CFG augmentation errors in DMD via a phased spatiotemporal schedule, eliminating redundant CFG computations and enabling single-pass inference per step; (3) Hybrid Long Tail Forcing, which enforces alignment exclusively on tail frames during long-horizon self-rollout training via a causal-bidirectional hybrid architecture, effectively mitigates spatiotemporal degradation in streaming mode while enhancing fidelity to reference frames; and (4) VAE Decoder Refiner through pixel-domain optimization of the VAE decoder to recover high-frequency details while circumventing latent-space ambiguities. Extensive experiments and analysis demonstrate that EchoTorrent achieves few-pass autoregressive generation with substantially extended temporal consistency, identity preservation, and audio-lip synchronization.
format Preprint
id arxiv_https___arxiv_org_abs_2602_13669
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle EchoTorrent: Towards Swift, Sustained, and Streaming Multi-Modal Video Generation
Meng, Rang
Wu, Weipeng
Li, Yuming
Ma, Chenguang
Computer Vision and Pattern Recognition
Recent multi-modal video generation models have achieved high visual quality, but their prohibitive latency and limited temporal stability hinder real-time deployment. Streaming inference exacerbates these issues, leading to pronounced multimodal degradation, such as spatial blurring, temporal drift, and lip desynchronization, which creates an unresolved efficiency-performance trade-off. To this end, we propose EchoTorrent, a novel schema with a fourfold design: (1) Multi-Teacher Training fine-tunes a pre-trained model on distinct preference domains to obtain specialized domain experts, which sequentially transfer domain-specific knowledge to a student model; (2) Adaptive CFG Calibration (ACC-DMD), which calibrates the audio CFG augmentation errors in DMD via a phased spatiotemporal schedule, eliminating redundant CFG computations and enabling single-pass inference per step; (3) Hybrid Long Tail Forcing, which enforces alignment exclusively on tail frames during long-horizon self-rollout training via a causal-bidirectional hybrid architecture, effectively mitigates spatiotemporal degradation in streaming mode while enhancing fidelity to reference frames; and (4) VAE Decoder Refiner through pixel-domain optimization of the VAE decoder to recover high-frequency details while circumventing latent-space ambiguities. Extensive experiments and analysis demonstrate that EchoTorrent achieves few-pass autoregressive generation with substantially extended temporal consistency, identity preservation, and audio-lip synchronization.
title EchoTorrent: Towards Swift, Sustained, and Streaming Multi-Modal Video Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2602.13669