The goal is to keep tract of incorrect classification when discovered.
This could include PR references, specific benchmarks, or input. Incorrect classifications should be stored in a permanent location. Such cases will be used to evaluate potential improvement in NVBench over time.
The goal is to keep tract of incorrect classification when discovered.
This could include PR references, specific benchmarks, or input. Incorrect classifications should be stored in a permanent location. Such cases will be used to evaluate potential improvement in NVBench over time.