The configuration and pre-trained model for Collective Activity Dataset

[tamtamz]

Thank you for the great work!
I tried to reproduce the result of Collective Activity Dataset but was able to achieve only about 88% accuracy.
If possible, can you provide the configuration file and pre-trained model for Collective Activity Dataset?

[Zxl1556180489]

File "/home/hi/datas/Conrad/GroupFormer-main/group/datasets/volleyball.py", line 257, in load_samples_sequence
flow = np.load(self.images_path + '/flow/%d/%d/%d.npy' % (sid, src_fid, fid))
FileNotFoundError: [Errno 2] No such file or directory: '/home/hi/data/volleyball/flow/37/25733/25730.npy'
Help！ How to get flow data？ thank you！