cannot extract model file from FlowNet2_checkpoint.pth.tar

[KaqiKaqi]

Hi, I downloaded the model file from chrome webbrowser directly and I got the FlowNet2_checkpoint.pth.tar file. Then I tried to extract the flownet2.pth from it as described in ./scripts/download_model.sh, but it turned out I couldn't extract it. I cannot use gdown to download the model file.

[Ahadiandian]

I also ran into the problem. You just need to rename the package flownet2.pth.tar to flownet2.pth. It's already a model file instead of a tar.

[jankais3r]

The Colab notebook is likely failing for the same reason.

[jiqibuxuexi]

Loading pretrained model from 'checkpoints/flownet2.pth'. Traceback (most recent call last): File "main.py", line 13, in dp.process(params) File "/home/gsk/consistent_depth-master/process.py", line 117, in process return self.pipeline(params) File "/home/gsk/consistent_depth-master/process.py", line 71, in pipeline self.flow.compute_flow(frame_pairs, params.flow_checkpoint) File "/home/gsk/consistent_depth-master/flow.py", line 143, in compute_flow optical_flow_flownet2_homography.process(args) File "/home/gsk/consistent_depth-master/optical_flow_flownet2_homography.py", line 250, in process flownet2_ckpt = torch.load(args.pretrained_model_flownet2) File "/usr/lib/python3/dist-packages/torch/serialization.py", line 419, in load f = open(f, 'rb') FileNotFoundError: [Errno 2] No such file or directory: 'checkpoints/flownet2.pth'

[jiqibuxuexi]

i also met this problem.it is hard for me to solve .damn it .

[ajinkyapuar]

I also ran into the problem. You just need to rename the package flownet2.pth.tar to flownet2.pth. It's already a model file instead of a tar.

Yes, this works.