Efficient Storage of Fine-Tuned Models via Low-Rank Approximation of Weight Residuals

Re-assign community