Dataset and model: Latent diffusion models for virtual battery material screening and characterization (Q6716001)

BattGen is a multimodal generative framework that accelerates virtual screening and characterization of battery materials. This framework uses latent diffusion model methodology to translate the data from characterization techniques such as atomic force microscopy to meaningful material information and screen battery materials based on battery functional properties such as average voltage, volume change, gravimetric and volumetric capacity, and working ion of required battery systems. This record contains the source data, model architecture, and training script used for the study. For access to machine learning tools defined in the training script, use https://gitlab.com/intelligent-analysis/cids/-/tree/v3.2a conda environment : requirements.txt