about dataset preparation

[Xutian-Ming]

Hi, thanks for your works！
I wonder could you please introduce the structure of directories in detail or introduce how to split train and val datasets in the form of json? (sorry that I don't know these so I asked)
I want to try DAIR-V2X dataset to infererence using the checkpoints you provided, but got the error of dataloading:

“ FileNotFoundError: [Errno 2] No such file or directory: 'dataset/my_dair_v2x/v2x_c/cooperative-vehicle-infrastructure/val.json' ”

and the original directory structure of DAIRV2X dataset is like:
├── my_dair_v2x
│ ├── v2x_c
label_world(many json files)
data_info.json
│ ├── v2x_i
calib
image
label
velodyne
data_info.json
│ └── v2x_v
calib
image
label
velodyne
data_info.json
There is no files like "val.json" so I want to ask you for help.

Hope for your reply, best wishes!

[yifanlu0227]

Hi! Please refer to step 3 in the website.

The annotations can be downloaded from the google drive

[Xutian-Ming]

Hi! Please refer to step 3 in the website. The annotations can be downloaded from the google drive

THANK YOU~

[Xutian-Ming]

Hi! Please refer to step 3 in the website. The annotations can be downloaded from the google drive

Sorry for another question, the downloadable DAIRV2X datasets will change every once in a while on the official website, so if it matters that I use different datasets instead of the dataset you used, thank you!

[yifanlu0227]

Will they change the content? I thought they are just trying keep these links alive.

I am not sure, I haven't logged into their site again once I downloaded the data.

[Xutian-Ming]

Will they change the content? I thought they are just trying keep these links alive.

I am not sure, I haven't logged into their site again once I downloaded the data.

yes, it will change. v_v
so I wonder if it is convenient to share the data you downloaded?