SLTrain: a sparse plus low rank approach for parameter and memory efficient pretrainingPublished in NeurIPS, 2024Download paper hereShare on Twitter Facebook LinkedIn Previous Next