sparseml icon indicating copy to clipboard operation
sparseml copied to clipboard

Published 20 hours ago • verified publisher icon neuralmagic

Adding HistogramObserver

Open abhinavnmagic opened this issue 1 year ago • 0 comments

The PR adds support for utilizing HistogramObserver from PyTorch which computes the min/max values for quantization by minimizing quantization error. The implementation has been tested on CodeLlama and Llama-2 models.

abhinavnmagic avatar Feb 08 '24 21:02 abhinavnmagic