LIM Component-Separation Dataset ([CII] + CO)

This dataset contains simulated line-intensity mapping (LIM) maps designed for machine-learning component separation tasks. Each sample consists of co-located 2D maps of total observed intensity, [CII] emission, CO emission, and instrumental noise at a single observed frequency (302 GHz), expressed in MJy/sr. The maps are derived from the SIDES-Uchuu simulations and include CONCERTO-inspired, frequency-dependent instrumental noise. The dataset is intended for training and benchmarking machine-learning models for spectral-line separation and denoising, and does not include beam convolution or angular correlations.