Are the weights of shufflenetv2.pth from ImageNet training ?

dog-qiuqiu / FastestDet

:zap: A newly designed ultra lightweight anchor free target detection algorithm， weight only 250K parameters， reduces the time consumption by 10% compared with yolo-fastest, and the post-processing is simpler

BSD 3-Clause "New" or "Revised" License

768 stars 131 forks source link

Are the weights of shufflenetv2.pth from ImageNet training ? #30

Open Hrayo712 opened 1 year ago

Hrayo712 commented 1 year ago

@dog-qiuqiu Are the shufflenetv2.pth weights which are loaded by default in shufflenetv2.py from ImageNet training ?

Hrayo712 commented 1 year ago

@nihui Perhaps you know ?

Hrayo712 commented 1 year ago

Any response from the developers ?