search

VitjanZ / 3DSR

Official implementation of the WACV2024 paper: Cheating Depth: Enhancing 3D Surface Anomaly Detection via Depth Simulation
Apache License 2.0
26 stars 0 forks source link

About Training Set Size #3

Open Leoliyao518 opened 3 months ago

Leoliyao518 commented 3 months ago

Hello！thanks for your excellent work. I want to know why the len() method of the dataset always returns 4000, but in fact most of the classes have only about 200 data in the training set. I tried to change the number of returns to the number of acquisitions, and deleted the code for randomly obtaining the index in the getitem method (keeping the same settings as the dataset of other methods), and the accuracy was greatly reduced.

SvyJ commented 5 days ago

I have also encountered the same problem, and I hope the author can answer it.