AIoT-MLSys-Lab / SVD-LLM

Official Code for "SVD-LLM: Truncation-aware Singular Value Decomposition for Large Language Model Compression"

https://arxiv.org/abs/2403.07378

AIoT-MLSys-Lab/SVD-LLM Issues

compresssed checkpoints
Updated a month ago1
still fail to apply on llama-13b
Closed 3 months ago2
fail to apply on llama-13b
Closed 3 months ago2
Incorrect Model Responses after compression
Closed 3 months ago4
Request for Code Integration of SVD-LLM with GPTQ
Closed 4 months ago2
recommended hardware resources
Closed 4 months ago1
Instead of ratio-compress to specific layer size
Closed 4 months ago1