GPU MODEL INFERENCE ACTIVE
Selected backend: onnxruntime_cuda
Runtime mode: GPU_INFERENCE_ACTIVE
This route is sealed with workflow navigation, usable actions, acceptance receipt, owner review, and public proof. High-impact actions remain owner-gated; no wallet, payment, mint, listing, payout, provider, or vault execution occurs here.
SCOS_PRODUCT_SURFACE_COMPLETE_ACCEPTANCE · externalActionExecuted=false