ICLR 2025 result

Our paper Mini-batch Coresets for Memory-efficient Language Model Training on Data Mixtures is accepted to ICLR 2025.