固定多模态框架，对比5种视觉主干在太阳能辐照预测中的表现。
A Controlled Visual-Backbone Benchmark for Multimodal Short-Term Solar Irradiance Forecasting
- 固定编码器、融合、损失等流程，仅更换视觉主干进行公平比较。
- 在Folsom数据集上，VMamba Small达65.39 W/m² RMSE，优于智能持久法。
- 结果可复现，适合想对比主干模型性能的研究者参考。
天空图像辐照度研究常同时改变图像编码器、时间模型、融合模块、目标定义和训练方案。本文采用更严格的协议：固定多模态预测流程，仅替换视觉主干。所有实验保持预处理、晴空指数归一化、气象历史编码、融合结构、回归头、损失函数、优化器调度、随机种子及时间划分策略一致。在Folsom和严格匹配的NREL数据集上，评估ConvNeXt、Swin Transformer、VMamba、Spatial Mamba和MambaVision五个主干模型在10分钟前预报中的表现。预报性能以晴空指数智能持久法为基准，仅报告时间模型作为气象历史诊断，不作为主要排名依据。在Folsom严格划分下，所有主干均优于智能持久法；单种子严格运行中，VMamba Small与Swin Base分别达到65.39和65.50 W/m² RMSE，而时间仅模型为69.51 W/m²。在313样本的NREL严格划分中，智能持久法仍最优（17.48 W/m²），最低视觉主干RMSE由Swin Tiny取得（23.76 W/m²）。该结果提供单一多模态配置下的可复现编码器对比，而非架构主导性结论或完全优化性能。代码已开源：https://github.com/Oshadha345/irradiance_benchmark
原文摘要 · Abstract (English)
Sky-image irradiance studies often compare forecasting systems in which the image encoder, temporal model, fusion block, target definition, and training recipe all change together. We use a narrower protocol: the multimodal forecasting pipeline is fixed, and only the visual backbone is varied. The shared setup keeps preprocessing, clear-sky-index normalization, weather-history encoding, fusion, regression head, loss, optimizer schedule, seed, and chronological split policy unchanged. We compare ConvNeXt, Swin Transformer, VMamba, Spatial Mamba, and MambaVision backbones for 10min-ahead forecasting on Folsom and a strict matched NREL split. Forecast skill is measured against clear-sky-index smart persistence, and temporal-only rows are reported as weather-history diagnostics rather than as the main ranking criterion. On the Folsom strict split, all evaluated visual-backbone runs improve over smart persistence. In the evaluated single-seed strict runs, VMamba Small and Swin Base reach matched Folsom RMSE values of 65.39 W/m^2 and 65.50 W/m^2; the temporal-only diagnostic reaches 69.51 W/m^2. On the 313-sample NREL strict split, smart persistence remains strongest at 17.48 W/m^2, while the lowest visual RMSE is obtained by Swin Tiny at 23.76 W/m^2. These results provide a reproducible encoder comparison under one fixed multimodal operating point rather than establishing architecture-level dominance, statistically resolved ranking, or fully optimized forecasting performance. Code available here: https://github.com/Oshadha345/irradiance_benchmark
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。