This storage project is used for deep-learning research at the signal processing group at the Institution for Electrical Engineering at Chalmers. The research consists of projects in, e.g., autonomous driving, semi-supervised learning, out-of-distribution detection, and medical imaging. The storage resources are used for storing datasets, model checkpoints, and results. Machine learning research conducted at the signal processing and computer vision group at electrical engineering, Chalmers In the groups, we are conducting research on deep neural network architectures and training schemes for detection, classification, and regression on a variety of sensory inputs, such as images, lidar point clouds, and time signals. The main task we will use Avlis for is to train deep neural networks using stochastic gradient descent on large datasets. We will typically develop and do initial experiments on local computers and deploy on Alvis for large-scale training which is impractical to perform on our desktop hardware. The training will typically be implemented using python-scripts contained in docker-containers. For our larger projects, training for one network could take months but our typical projects are more on the order of days. We use both public data such as KITTI, NuSence. Lyft 5 etc. as well as our own datasets available at visuallocalization.net.