The video names in Split do not match the actual processed data.

[YoungjaeDev]

https://github.com/NVIDIA-AI-IOT/tao_toolkit_recipes/blob/cdd7871bb765082facace33c02b6f8a33a79e30f/tao_action_recognition/data_generation/split_dataset.py#L38

I am receiving data from here and preprocessing the action recognition data, but there is no match in the dataset.

!wget -P $HOST_DATA_DIR http://serre-lab.clps.brown.edu/wp-content/uploads/2013/10/hmdb51_org.rar

https://catalog.ngc.nvidia.com/orgs/nvidia/teams/tao/resources/tao-getting-started

[YoungjaeDev]

https://catalog.ngc.nvidia.com/orgs/nvidia/teams/tao/resources/cv_samples

hmdb51 is still the same version in the file here, can you please check?

[Tyler-D]

Do you mean the video clip names in CLASS_test_split1.txt do not match with the name in the datastet ? Can you give a sample of such mismatch ?

[YoungjaeDev]

@Tyler-D For example, a command like this: Of course, the path to the file is correct, and there is a video like bad_53, not bad_21.

Traceback (most recent call last):
  File "/home/aicads/anaconda3/envs/launcher/lib/python3.6/shutil.py", line 550, in move
    os.rename(src, real_dst)
FileNotFoundError: [Errno 2] No such file or directory: '/home/aicads/workspace/tao-getting-started_v5.0.0/notebooks/tao_launcher_starter_kit/action_recognition_net/data/actionrecognitionnet/processed_data/fall_floor/BIG_FISH_fall_floor_f_nm_np1_fr_bad_21' -> '/home/aicads/workspace/tao-getting-started_v5.0.0/notebooks/tao_launcher_starter_kit/action_recognition_net/data/actionrecognitionnet/train/fall_floor/BIG_FISH_fall_floor_f_nm_np1_fr_bad_21'

During handling of the above exception, another exception occurred:

[YoungjaeDev]

Don't you?

[Tyler-D]

I can see a video named BIG_FISH_fall_floor_f_nm_np1_fr_bad_21.avi in the hmdb51_org.rar I downloaded (re-downloaded just know). And there is no such video with bad_53 in the fall_floor directory.

[YoungjaeDev]

bad_53 is simply example... I'll try again tomorrow

[YoungjaeDev]

It's fine now. But, I don't know why not works... lol