r38: ppo_cho2 chain (choose wire 0.800), ft_skill5_ppo (stage-2 skill heads), refreshed card measurements

Upload folder using huggingface_hub

Card: architecture as rendered PNG (hub does not render mermaid)

Lean release: model + card + run.py. Reproduction lives in the GitHub repo (make model). Plugin: github.com/Andrei-cloud/laya-plugin

policy release 20/76