Video-Panda: Parameter-efficient Alignment for Encoder-free Video-Language Models

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Yi, Jinhui, Wasim, Syed Talal, Luo, Yanan, Naseer, Muzammal, Gall, Juergen
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908286484217856
author Yi, Jinhui
Wasim, Syed Talal
Luo, Yanan
Naseer, Muzammal
Gall, Juergen
author_facet Yi, Jinhui
Wasim, Syed Talal
Luo, Yanan
Naseer, Muzammal
Gall, Juergen
contents We present an efficient encoder-free approach for video-language understanding that achieves competitive performance while significantly reducing computational overhead. Current video-language models typically rely on heavyweight image encoders (300M-1.1B parameters) or video encoders (1B-1.4B parameters), creating a substantial computational burden when processing multi-frame videos. Our method introduces a novel Spatio-Temporal Alignment Block (STAB) that directly processes video inputs without requiring pre-trained encoders while using only 45M parameters for visual processing - at least a 6.5$\times$ reduction compared to traditional approaches. The STAB architecture combines Local Spatio-Temporal Encoding for fine-grained feature extraction, efficient spatial downsampling through learned attention and separate mechanisms for modeling frame-level and video-level relationships. Our model achieves comparable or superior performance to encoder-based approaches for open-ended video question answering on standard benchmarks. The fine-grained video question-answering evaluation demonstrates our model's effectiveness, outperforming the encoder-based approaches Video-ChatGPT and Video-LLaVA in key aspects like correctness and temporal understanding. Extensive ablation studies validate our architectural choices and demonstrate the effectiveness of our spatio-temporal modeling approach while achieving 3-4$\times$ faster processing speeds than previous methods. Code is available at https://jh-yi.github.io/Video-Panda.
format Preprint
id arxiv_https___arxiv_org_abs_2412_18609
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Video-Panda: Parameter-efficient Alignment for Encoder-free Video-Language Models
Yi, Jinhui
Wasim, Syed Talal
Luo, Yanan
Naseer, Muzammal
Gall, Juergen
Computer Vision and Pattern Recognition
We present an efficient encoder-free approach for video-language understanding that achieves competitive performance while significantly reducing computational overhead. Current video-language models typically rely on heavyweight image encoders (300M-1.1B parameters) or video encoders (1B-1.4B parameters), creating a substantial computational burden when processing multi-frame videos. Our method introduces a novel Spatio-Temporal Alignment Block (STAB) that directly processes video inputs without requiring pre-trained encoders while using only 45M parameters for visual processing - at least a 6.5$\times$ reduction compared to traditional approaches. The STAB architecture combines Local Spatio-Temporal Encoding for fine-grained feature extraction, efficient spatial downsampling through learned attention and separate mechanisms for modeling frame-level and video-level relationships. Our model achieves comparable or superior performance to encoder-based approaches for open-ended video question answering on standard benchmarks. The fine-grained video question-answering evaluation demonstrates our model's effectiveness, outperforming the encoder-based approaches Video-ChatGPT and Video-LLaVA in key aspects like correctness and temporal understanding. Extensive ablation studies validate our architectural choices and demonstrate the effectiveness of our spatio-temporal modeling approach while achieving 3-4$\times$ faster processing speeds than previous methods. Code is available at https://jh-yi.github.io/Video-Panda.
title Video-Panda: Parameter-efficient Alignment for Encoder-free Video-Language Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.18609