ekosman/AnomalyDetectionCVPR2018-Pytorch

AttributeError: 'numpy.ndarray' object has no attribute 'cpu'

Closed this issue · 3 comments

Need some help in solving this issue
while running "feature_extractor" I get this error:

     36                         for i, (dir, vid_name, start_frame) in enumerate(zip(dirs, vid_names, sampled_idx.cpu().numpy())):
     37                                 dir = path.join(save_dir, dir)
---> 38                                 features_writer.write(feature=outputs[i], video_name=vid_name, idx=start_frame, dir=dir)
     39 
     40         features_writer.dump()

<ipython-input-7-563b8b257603> in write(self, feature, video_name, idx, dir)
     48                         self._init_video(video_name, dir)
     49 
---> 50                 self.store(feature, idx)

<ipython-input-7-563b8b257603> in store(self, feature, idx)
     38 
     39         def store(self, feature, idx):
---> 40                 self.data[idx] = list(feature.cpu().numpy())
     41 
     42         def write(self, feature, video_name, idx, dir):

AttributeError: 'numpy.ndarray' object has no attribute 'cpu'

Anyone facing the same issue? Or if anyone who has solved this can help

I think I have had this before but can't remember why.
Could you give me the "type()" of the feature variable? maybe from just before the call at 40?

Thank you I solved this error
We can close this issue

Thank you I solved this error We can close this issue

how?thank you