We provide an open-source Langevin field-theoretic simulation (L-FTS) for the block copolymer (BCP) field-based simulations. L-FTS is one of the partial saddle point approximation methods that can account for the compositional fluctuation effect, which is neglected in the self-consistent field theory (SCFT). By parallelizing the computations using the GPU, we achieved a 20x speedup in typical BCP systems compared to the CPU implementation. Our open-source code is provided as a library for Python programming language, so that one can implement SCFT calculation and L-FTS by writing a Python script. This allows L-FTS to be integrated with numerous useful Python libraries. We implemented deep learning boost of the L-FTS on top of this library and is also publicly available, which provides additional 6x speedup without compromising accuracy. *This research was supported by the NRF grants (NRF-2021R1A2C1011072, and NRF-2022R1C1C2010613) funded by the Ministry of Science and ICT (MSIT), Korea. This research used high-performance computing resources of the UNIST Supercomputing Center.