Missing files './data/example_video', 'visual_util.py' and 'default_dataset.yaml'

[DengKaiCQ]

Hi, authors

In the file ./training/config/omniworld_full.yaml there is a required base yaml

defaults:
  - default_dataset.yaml
 ...

but we can't find it in the training/config

MoRe $ cd training/config
config $ ls
omniworld_full.yaml  omniworld_stream.yaml
config $ tree
.
├── omniworld_full.yaml
└── omniworld_stream.yaml

1 directory, 2 files

Also we can't find the ./data/example_video like this

MoRe $ ls
eval        inference.py  preprocess  __pycache__  requirements-noversion.txt  training  viser
index.html  media         pretrained  README.md    requirements.txt            vggt      visual_util.py

The visual_util.py is copied from VGGT and it is also missing.

And when running the following cmd, below are what I got

MoRe $ python inference.py \
    --config_path training/config/omniworld_full.yaml \
    --ckpt_path pretrained/more_full.pt \
    --image_path ./data/example_video \
    --output_dir ./results/full_res \
    --conf_thres 50.0 \
    --predict_motion
[*] Loading model from pretrained/more_full.pt...
.../python3.11/site-packages/hydra/_internal/defaults_list.py:251: UserWarning: In 'omniworld_full': Defaults list is missing `_self_`. See https://hydra.cc/docs/1.2/upgrades/1.0_to_1.1/default_composition_order for more information
  warnings.warn(msg, UserWarning)
[Error] In 'omniworld_full': Could not load 'default_dataset.yaml'.

MoRe $ python inference.py \
    --config_path training/config/omniworld_full.yaml 
    --ckpt_path pretrained/more_full.pt  \
   --image_path .../Pi3/examples/parkour_long.mp4 \
    --output_dir ./results/full_res  \
   --conf_thres 50.0  \
   --predict_motion
[*] Loading model from pretrained/more_full.pt...
[Error] Error locating target 'vggt.models.vggt.VGGT', set env var HYDRA_FULL_ERROR=1 to see chained exception.
full_key: model

[Wendyiii]

Hi, in VGGT GitHub Repo, you can find these two files.