Researchers from ETH Zurich, EPFL, and Microsoft Introduce QuaRot: A Machine Learning Method that Enables 4-bit Inference of LLMs by Removing the Outlier Features – MarkTechPost
Researchers from ETH Zurich, EPFL, and Microsoft Introduce QuaRot: A Machine Learning Method that Enables 4-bit Inference of LLMs by Removing the Outlier Features – MarkTechPost