BenchmarkMixed-WM38 Dataset

Mixed-WM38

A semiconductor defect classification dataset containing 6,866 wafer maps, used to evaluate zero-shot wafer map classification methods.

Timeline (last 90 days)

Oct 27

On the Mixed-WM38 dataset (containing 6,866 wafer maps), the multi-region prompting strategy achieved 92.9% macro-average accuracy and 93.8% overall accuracy.

Unverified50%

All Facts (1)

Unverified

On the Mixed-WM38 dataset (containing 6,866 wafer maps), the multi-region prompting strategy achieved 92.9% macro-average accuracy and 93.8% overall accuracy.

50%

Source Articles

Zero-Shot Wafer Map Classification: A Training-Free Multimodal LLM ApproachOct 27