Single diverging wave (DW) imaging produces ultrasound (US) images at high frame rate (ultrafast) but of low quality. Conventional high-quality DW imaging relies on the coherent compounding of multiple consecutive steered emissions, which in turn reduces the gain in frame rate. Reconstructing high-quality US images for ultrafast imaging using deep learning techniques has recently raised a growing interest in the US community. We recently described a convolutional neural network (CNN) architecture called ID-Net, which exploited an inception layer devoted to the reconstruction of DW ultrasound images using radio frequency (RF) data. We derive in this work the complex equivalent of this network, i.e., the complex inception for DW network (CID-Net), operating on in-phase/quadrature (I/Q) data. We experimentally demonstrate that the CID-Net yields the same image quality as that obtained from the RF-trained CNN, i.e., using only three I/Q images, the CID-Net yields high-quality images competing with those obtained by coherently compounding 31 RF images