@article{towardsmemoryefficienttrainingfor, title = {Towards Memory-Efficient Training for Extremely Large Output Spaces -- Learning with 500k Labels on a Single Commodity GPU}, author = {Erik Schultheis and Rohit Babbar}, year = {2023}, eprint = {2306.03725}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2306.03725v1}, }