TransiNet (version 2) is the umbrella name for the official project wrapping around multiple deep learning-based transient detection implementations for the LSST telescope — a.k.a. the Vera C. Rubin Observatory.
The original TransiNet (TransiNet0) is an end-to-end image-generating astronomical transient detection that performs de-noising, PSF-matching, localization and detection all in a single go.
For LSST, however, any deep learning-based implementation with the aim of transient detection has the term TransiNet in its name. So, although an implementation based on the original idea of TransiNet0 is underway, the traditional real-bogus classifier is also called the RBTransiNet.