The TanDEM-X Forest/Non-Forest map, derived from the volume decorrelation factor using a supervised fuzzy clustering algorithm, represents the baseline approach for forest mapping with TanDEM-X data at global scale. Deep learning (DL) methods have been demonstrated to be also suitable for mapping forests at large scale with TanDEM-X interferometric data. In this work, we investigate the capabilities of using a U-Net-like architecture with TanDEM-X interferometric data for forest mapping at 6 m resolution. With such high-resolution data, we aim at improving the forest mapping accuracy and to be able to detect forest degradation over the Amazon rainforest caused e.g. by selective logging, fires and natural hazards. The classification improvements already observed applying DL methods on TanDEM-X data allow for the generation of large scale time-tagged mosaics. The explotation of such mosaics over extended areas is a key aspect for the detection and monitoring of forest dynamics worldwide.