ProcrustesGPT: Compressing LLMs with Structured Matrices and Orthogonal Transformations

Open in new window