3D-ResNets-PyTorch
3D-ResNets-PyTorch copied to clipboard
Output of hmdb51_json.py
After evaluation we get 3 hmdb51_{}.json files, but for training we can use only one annotation fine. What a point of split_index in hmdb51_json.py? Tran model with a hmdb51_1.json and it's going great, but what a point of 2 another outputs?
What did you mean in the question?
I also have this question that why divide hmdb51 into 3 json files.
I also have this question that why divide hmdb51 into 3 json files.and i don't how to convert a json like activity by using hmdb51,if someone solve this problem ,pleaase contact me