… For this exercise, I chose Google Gemma 4 31B UD-Q8 K XL , a near-lossless quantization, somewhat ambitiously, as I will learn shortly. …
… Because I’m running multiple editor and reporter processes in parallel, the parameter count on each model may seem undersized for the 96GB of RAM that my AMD GPU can access after some BIOS tweaks : I’m using a mix of Qwen’s straightforward 3.5-9B model, as well as Jackrong’s Qwen-3.5-9B-GLM-5.1-Dis… …
… In the HDR10 mode, there are three Perceptual Quantization PQ options, Clip, Basic and Optimized. …