Automated planning is a core problem in artificial intelligence, where the goal is to compute sequences of actions that achieve specified goals in structured, symbolic environments. Recent work has shown that deep learning models, in particular transformer architectures, can be trained to generate plans or guide search. The purpose of this project is to investigate the usage of deep learning models for automated planning tasks. The project requires GPU resources to train transformer models on extensive planning datasets and to perform systematic experiments across multiple planning domains and problem sizes. The results are expected to advance understanding of machine learning-based planning research.