Difficulty understanding what should be passed as --pose and --res arguments for step 5 of data preprocessing

So I have created a toy dataset (since step 4 is taking a lot of time for me and I want to for now just make sure I have an understanding of the entire platform).

So, my toy dataset has a train and test section sampled from HICO-DET dataset.

How should I pass arguments to --pose and --res?

Could you please help?

The following is shown in the README step 5: $ python script/assign_pose_GT.py --pose <path to your pose used for SMPLify-X> --res <path to your SMPLify-X result>

so should I do the following?

$ python script/assign_pose_GT.py --pose /scratch3/research/data/smplify-x/ride_bike/toy_dataset/tmp_data_train/keypoints --res /scratch3/research/data/smplify-x/ride_bike/toy_dataset/tmp_res_train/results

Which yields to Data/Trainval_GT_HICO_with_idx.pkl file and only for train split or should I mix train and test data?

Could you please show the following command with the exact arguments you pass to it for --pose and --res? $ python script/assign_pose_GT.py --pose <path to your pose used for SMPLify-X> --res <path to your SMPLify-X result>

Here's my toy_dataset structure:

[jalal@goku ride_bike]$ tree toy_dataset/
toy_dataset/
├── tmp_data_test
│   ├── images
│   │   ├── HICO_test2015_00000044.jpg
│   │   └── HICO_test2015_00000061.jpg
│   └── keypoints
│       ├── HICO_test2015_00000044_keypoints.json
│       └── HICO_test2015_00000061_keypoints.json
├── tmp_data_train
│   ├── images
│   │   ├── HICO_train2015_00000015.jpg
│   │   ├── HICO_train2015_00000019.jpg
│   │   ├── HICO_train2015_00000060.jpg
│   │   ├── HICO_train2015_00000140.jpg
│   │   ├── HICO_train2015_00000202.jpg
│   │   └── HICO_train2015_00000232.jpg
│   └── keypoints
│       ├── HICO_train2015_00000015_keypoints.json
│       ├── HICO_train2015_00000019_keypoints.json
│       ├── HICO_train2015_00000060_keypoints.json
│       ├── HICO_train2015_00000140_keypoints.json
│       ├── HICO_train2015_00000202_keypoints.json
│       └── HICO_train2015_00000232_keypoints.json
├── tmp_res_test
│   ├── conf.yaml
│   ├── images
│   │   ├── HICO_test2015_00000044
│   │   │   └── 000
│   │   └── HICO_test2015_00000061
│   │       └── 000
│   ├── meshes
│   │   ├── HICO_test2015_00000044
│   │   │   └── 000.obj
│   │   └── HICO_test2015_00000061
│   │       └── 000.obj
│   └── results
│       ├── HICO_test2015_00000044
│       │   └── 000.pkl
│       └── HICO_test2015_00000061
│           └── 000.pkl
└── tmp_res_train
    ├── conf.yaml
    ├── images
    │   ├── HICO_train2015_00000015
    │   │   └── 000
    │   ├── HICO_train2015_00000019
    │   │   └── 000
    │   ├── HICO_train2015_00000060
    │   │   ├── 000
    │   │   ├── 001
    │   │   └── 002
    │   ├── HICO_train2015_00000140
    │   │   └── 000
    │   ├── HICO_train2015_00000202
    │   │   └── 000
    │   └── HICO_train2015_00000232
    │       ├── 000
    │       └── 001
    ├── meshes
    │   ├── HICO_train2015_00000015
    │   │   └── 000.obj
    │   ├── HICO_train2015_00000019
    │   │   └── 000.obj
    │   ├── HICO_train2015_00000060
    │   │   ├── 000.obj
    │   │   ├── 001.obj
    │   │   └── 002.obj
    │   ├── HICO_train2015_00000140
    │   │   └── 000.obj
    │   ├── HICO_train2015_00000202
    │   │   └── 000.obj
    │   └── HICO_train2015_00000232
    │       └── 000.obj
    └── results
        ├── HICO_train2015_00000015
        │   └── 000.pkl
        ├── HICO_train2015_00000019
        │   └── 000.pkl
        ├── HICO_train2015_00000060
        │   ├── 000.pkl
        │   ├── 001.pkl
        │   └── 002.pkl
        ├── HICO_train2015_00000140
        │   └── 000.pkl
        ├── HICO_train2015_00000202
        │   └── 000.pkl
        └── HICO_train2015_00000232
            └── 000.pkl

49 directories, 38 files

It is not really clear to me which part of data (train vs test) is passed to each of the 3 commands below. Could you please clarify with an example?

Screenshot from 2021-05-26 18-22-26

DirtyHarryLYL / DJ-RN

Difficulty understanding what should be passed as --pose and --res arguments for step 5 of data preprocessing #54