Radar‑Echo‑Nowcasting‑CNN‑LSTM

基于CNN‑LSTM的雷达回波外推短时临近预报预训练权重。

模型信息

  • 架构:CNN‑LSTM 时空预测模型
  • 训练硬件:RTX4060 8G
  • 输入:10帧 256×256 灰度雷达图像
  • 输出:预测未来10帧雷达回波
  • 权重文件:best_model.pth (3.26GB)

验证集指标

MSE MAE SSIM CSI@0.184 CSI@0.3
0.0080 0.0566 0.7562 0.5606 0.1614

使用方式

完整源代码、训练/推理脚本:https://github.com/itcallsitcalls/Radar-Echo-Nowcasting-CNN-LSTM

推理示例

下载本仓库best_model.pth,配合GitHub项目源码运行:

python src/inference.py --best_model.pth
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support