根据输入动态剪枝模型与数据，实现高效时空预测。
DYNA-PRUNER: Input-Adaptive Data-Model Co-Pruning for Efficient and Scalable Spatio-Temporal Media Prediction
- 输入自适应联合剪枝数据与模型结构，生成稀疏子网络。
- 在多个数据集上减少70%计算量，速度提升2.5倍，精度损失小于1%。
- 适合资源受限场景下的实时时空媒体分析，兼容主流模型架构。
时空预测支持雷达/卫星短时预报和城市级交通监控，但现代模型常因计算密集难以实时部署。其根源在于密集计算与强输入依赖冗余（如平静海面或晴朗天空）之间的不匹配。为实现可扩展媒体分析中自动化的、资源感知的架构优化，我们提出 Dyna-Pruner，一个端到端的输入依赖型数据-模型联合剪枝框架。通过共享重要性同步机制生成耦合掩码，剪除冗余区域及其对应计算单元（如卷积滤波器），在推理时生成每样本的稀疏子网络。在 WeatherBench、SEVIR 和 TaxiBJ 上的实验表明，该方法可无缝集成于 CNN、RNN 及 Transformer 骨干网络，在保持精度损失低于 1% 的前提下，最多减少 70% 的浮点运算量，并在 NVIDIA Jetson AGX Orin 上实现 2.5 倍的速度提升。
原文摘要 · Abstract (English)
Spatio-temporal prediction supports radar/satellite nowcasting and city-scale traffic monitoring, but modern models are often too expensive for real-time deployment. This stems from a mismatch between dense computation and strong input-dependent redundancy (e.g., calm seas or clear skies). To enable automated, resource-aware architecture optimization in scalable media analysis, we propose Dyna-Pruner, an end-to-end framework for input-dependent co-pruning of data and model structure. A shared-importance synchronization mechanism generates coupled masks that prune redundant regions and their corresponding computational units (e.g., convolutional filters), yielding per-sample sparse sub-networks at inference time. Experiments on WeatherBench, SEVIR, and TaxiBJ show seamless integration with CNN, RNN, and Transformer backbones, reducing FLOPs by up to $70\%$ and achieving a $2.5\times$ speedup on NVIDIA Jetson AGX Orin with negligible accuracy loss ($<1\%$).
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。