The objective of this internship is to set up a diffusion model on a physical database such as images of physical measurements. We seek the distributions of images conditionally on physical parameters. Since the physics simulation code is expensive (several hours on a supercomputer), the database is small. Another difficulty lies in the fact that the processed images are maps of physical quantities and will require calibration. (In machine learning, images are usually normalized). This work could continue in thesis by developing the resolution of an inverse problem using a diffusion model. In accordance with the commitments made by the CEA in favor of the integration of people with disabilities, this job is open to all and all. Participating in national protection, an administrative investigation is carried out for all CEA employees in order to ensure the integrity and security of the nation.