The 4m International Liquid Mirror Telescope (ILMT) is India’s first optical survey telescope, which uses a spinning mercury mirror as its primary reflector. It is the only operational liquid mirror telescope globally, with a 22.3' FoV and limiting magnitude of g-mag ~22. The ILMT captures ~15 GB of imaging data per night, producing nearly 30 science frames. A key goal is discovering optical transients and variables, which is facilitated by an automated detection pipeline called the PyLMT. The pipeline employs image subtraction to generate difference frames, followed by a CNN-based classifier to distinguish real sources from artefacts in the resulting images. A second CNN classifies real sources based on host morphology, and known asteroids and solar system objects are identified for rejection. In a recent modification to the pipeline, new CNN models for transient detection and classification were trained using the transfer learning technique, employed between images from the Zwicky Transient Facility and the ILMT. Operational since November 2023, the pipeline has identified various transients and variables, including AGNs, asteroids, T Tauri outbursts, RR Lyrae, Eclipsing Binaries, and 21 supernova-like transient candidates. Five of the detected supernova candidates, viz. AT 2023yjc, 2024fxn, 2024zsm, 2024agkc, and 2024aifv were reported to the transient name server (TNS) as new discoveries.