search

StevenLiuWen / ano_pred_cvpr2018

Official implementation of Paper Future Frame Prediction for Anomaly Detection -- A New Baseline, CVPR 2018
435 stars 126 forks source link

training error about flownet #8

Closed shamoqianting closed 5 years ago

shamoqianting commented 6 years ago

Dear Dr. LiuWen, I am trying to train from scratch following your instruction. But I encountered the following error:

Traceback (most recent call last): File "train.py", line 94, in <module> height=flow_height, width=flow_width, reuse=None) File "/data1/ShanghaiTech-FutureFramePred/ano_pred_cvpr2018-master/Codes/models.py", line 25, in flownet net = FlowNetSD(mode=Mode.TEST) TypeError: Can't instantiate abstract class FlowNetSD with abstract methods loss

Could you help to check this ? Thank you very much

shamoqianting commented 6 years ago

It seems that the error caused by a missing method "loss" which should be overridden in Codes/flownet2/src/flownet_sd/flownet_sd.py . I just uncommentted the loss function and the program began to run, although I am not sure the result is correct or not.