A feature-compressed multi-task learning U-Net for shallow-water source localization in the presence of internal waves

•A feature-compressed multi-task learning U-Net with CBAM (MTL-UNET-CBAM) for shallow-water source localization with the presence of internal waves is constructed.•CBAM improves range localization performance of MTL-UNET, but has no significant improvement on depth estimation.•MTL-UNET-CBAM has higher localization robustness and computation speed then CMFP.•Applying DTL to MTL-UNET-CBAM significantly improve the localization performance of MTL-UNET-CBAM on both range and depth estimation.