DISP-LLM: Dimension-Independent Structural Pruning for Large Language Models

Structural pruning has emerged as a promising solution to reduce the costs of LLMs without requiring post-processing steps.