Question about training steps used to achieve reported performance

nelson1425 / EfficientAD

Unofficial implementation of EfficientAD https://arxiv.org/abs/2303.14535

https://arxiv.org/abs/2303.14535

Apache License 2.0

273 stars 72 forks source link

Open willyfh opened 7 months ago

willyfh commented 7 months ago

For the performance reported in the readme, I wonder, did you use the same training steps (70,000) for all datasets (mvtec, mvtec loco, and visa)?