Compilation of a Petri net model is one way to accelerate its veri cation through state space exploration. In this approach, code to explore the Petri net is generated, which avoids the use a xed exploration tool involving an interpretation" of the Petri net structure. In this paper, we show how peculiarities in the model (like places types, boundedness, invariants, etc., known by construction if adequate modelling tools are used) can be exploited to improve the generated code e ciency. We present a lightweight code generation framework targeting the LLVM language and demonstrate its exibility by considering several kinds of optimisations. Our compiler is called lightweight because the annotations of the compiled models are directly expressed in the target language, which saves from translating them and allows for fast compilation. The accelerations resulting from the optimisations are then evaluated on various Petri net models, showing important speedups and in some cases outperforming state-of-the-art tools.