Add support section to model card

Quick Start: gated-base prerequisite note; clarify total download is unchanged vs base

Move whole-pipeline VRAM table into the top overview next to the encoder comparison

Benchmarks: measured whole-pipeline VRAM tiers (co-resident vs encoder-offloaded, 1024/768); clarify encoder-only VRAM row

Initial release: pruned 5.1B text encoder for FLUX.2-klein-9B (bitwise-verified export, benchmarks, seed study)