Recent advancements in machine learning models utilizing Hamiltonian formulations have shown promising results for both energy-conserving and non-energy-conserving simple mechanical systems. In this talk, we consider a generalization of the Hamiltonian formulation, and present pseudo-Hamiltonian neural networks (PHNN) for learning partial differential equations. This model is comprised of up to three neural networks, modelling terms representing conservation, dissipation and external forces, and discrete convolution operators that can either be learned or be prior knowledge. We demonstrate numerically the superior performance of PHNN compared to a baseline model that models the full dynamics by a single neural network, and discuss the advantages of having a model that can be separated into terms with different physical interpretations.