7. Key Quality Metrics

Performance-related metrics must remain key acceptance criteria throughout Stages 3-6. Stage 3 focuses on operator performance, and Stage 4 focuses on inference accuracy and end-to-end performance.

Stage	Acceptance criteria	Pass standard	Required hardware
Stage 0	MOU signing	Effective after both parties sign or seal	N/A
Stage 1	Environment readiness	Chip is recognized by the OS and PyTorch can recognize the backend	1 development board
Stage 2	Triton kernel compilation pass rate	>=95%, CI machine availability >=99%	1 CI machine
Stage 3	FlagGems operator test pass rate	>=80%, core operators 100%	2 CI machines and 1 debug machine
Stage 4	Model accuracy deviation	<2% relative to the NVIDIA baseline	2 CI machines and 1 debug machine
Stage 5	CTS core tests passed	100% passed or waiver approved	2 CI machines and 2 SVT machines
Stage 6	Number of models released through FlagRelease	>=10 models	1 dedicated FR machine