CoMERA: Computing- and Memory-Efficient Training via Rank-Adaptive Tensor Optimization

VLM
Re-assign community