ML-TANGO / TANGO

public repo for TANGO (Target Aware No-code neural network Generation and Operation framework)
Other
68 stars 20 forks source link

YoloE: evaluation batch size is same as the one of train #87

Closed iksooman closed 11 months ago

iksooman commented 11 months ago

YoloE가 테스트 수행 시에 batch size를 train batch size의 2배를 하는 방식에서 train의 batch size를 그대로 사용하는 방식으로 변경되었습니다.