model card: standard structure, measured facts, verified license declarations

gpu-pipelined: add the missing tokenizer/generation_config.json (fix by addition; byte-identical to the file hybrid-pf64/ ships from the same snapshot)

README: hybrid-pf64 bundles

Add hybrid-pf64/gemma4_31b_qat_decode_int4linsym_msdpa_g8_pf64

docs: expected shape-specialization memory-growth note (sibling-measured, unverified here)

docs: refresh measured performance (local retest 2026-08-18)

docs: descriptive model card

Add gemma4_31b_qat_decode_int4linsym_msdpa_g8

Upload README.md with huggingface_hub

initial commit