How to download nuScenes-C dataset? I did not find the relevant data set in this link https://opendatalab.com/OpenDataLab/nuScenes-C.Thanks!

Hi, is there any way I can plot the results demo on RoboBEV based on my modal detection results?

when I use 'bash tools/generate_dataset.sh', it raised an error:
img file does not exist: ./data/nuscenes/samples/CAM_FRONT/n015-2018-07-11-11-54-16+0800__CAM_FRONT__1531281439762460.jpg
Actually I do have this file in my path. And it seems like find another file path in my mmdet3d location. It still use loading.py from mmdet3d rather than custom pipeline

But I have already changed the config file dataset path and changed the PYTHONPATH in generate_dataset.sh. How can I fix it?

1.run the command bash /uda/tools/create_data.sh

The content of the create_data.sh file is as follows:

#!/usr/bin/env bash

PYTHONPATH="$(dirname $0)/..":$PYTHONPATH \
python /home/yangxue/code/RoboBEV/uda/tools/create_data.py \
nuscenes \
--root-path /home/yangxue/code/RoboBEV/data/nuscenes \
--out-dir '/home/yangxue/code/RoboBEV/data/uda_new' \
--version 'v1.0-mini' \
--domain 'city2city' \
--extra-tag nuscenes \
--canbus /home/yangxue/code/RoboBEV/data/nuscenes/

The file structure in the'/home/yangxue/code/RoboBEV/data/nuscenes' folder is as follows:

/home/yangxue/code/RoboBEV/data/nuscenes/can_bus
/home/yangxue/code/RoboBEV/data/nuscenes/lidarseg
/home/yangxue/code/RoboBEV/data/nuscenes/maps
/home/yangxue/code/RoboBEV/data/nuscenes/samples
/home/yangxue/code/RoboBEV/data/nuscenes/sweeps
/home/yangxue/code/RoboBEV/data/nuscenes/v1.0-mini
/home/yangxue/code/RoboBEV/data/nuscenes/v1.0-mini-test
/home/yangxue/code/RoboBEV/data/nuscenes/v1.0-mini-trainval
/home/yangxue/code/RoboBEV/data/nuscenes/nuscenes_infos_temporal_val.pkl

The file structure in the'/home/yangxue/code/RoboBEV/data/uda_new' folder is as follows:

nuscenes_infos_boston_train.pkl
nuscenes_infos_boston_train_mono3d.coco.json
nuscenes_infos_boston_val.pkl
nuscenes_infos_boston_val_mono3d.coco.json
nuscenes_infos_sing_train.pkl
nuscenes_infos_sing_train_mono3d.coco.json
nuscenes_infos_sing_val.pkl
nuscenes_infos_sing_val_mono3d.coco.json

2.run the command bash /corruptions/tools/generate_dataset.sh

The content of the generate_dataset.sh file is as follows:

#!/usr/bin/env bash

CONFIG=/home/yangxue/code/RoboBEV/corruptions/project/config/nuscenes_c.py

PYTHONPATH="$(dirname $0)/..":$PYTHONPATH \
python $(dirname "$0")/generate_dataset.py $CONFIG

The content of the nuscenes_c.py file is as follows:

data_root = '/home/yangxue/code/RoboBEV/data/nuscenes/'
anno_root = '/home/yangxue/code/RoboBEV/data/nuscenes/'
corruption_root = '/home/yangxue/code/RoboBEV/data/nuScenes-c/'

data = dict(
    samples_per_gpu=1,
    workers_per_gpu=16,
    val=dict(type=dataset_type,
             data_root=data_root,
             ann_file=anno_root + 'nuscenes_infos_temporal_val.pkl',
             pipeline=test_pipeline,
             classes=class_names, modality=input_modality),
    shuffler_sampler=dict(type='DistributedGroupSampler')

The following error is reported in step 2:

FileNotFoundError: img file does not exist: ./data/nuscenes/samples/CAM_FRONT/n008-2018-08-01-15-16-36-0400__CAM_FRONT__1533151603512404.jpg

For the occupancy task, the data and evaluation used for the cvpr2023 competition (https://github.com/CVPR2023-3D-Occupancy-Prediction/CVPR2023-3D-Occupancy-Prediction) are more popular, does this repository have plans to use it for the robust occupancy task?

Thank you for your excellent work. When trying to generate the nuscnes-C dataset, I met an error that

Traceback (most recent call last):
  File "tools/generate_dataset.py", line 127, in <module>
    main()
  File "tools/generate_dataset.py", line 97, in main
    test = dataset[0]
  File "/workspace/RoboBEV/corruptions/project/mmdet3d_plugin/datasets/custom_nuscenes_dataset.py", line 176, in __getitem__
    return self.prepare_test_data(idx)
  File "/workspace/mmdetection3d/mmdet3d/datasets/custom_3d.py", line 245, in prepare_test_data
    input_dict = self.get_data_info(index)
  File "/workspace/RoboBEV/corruptions/project/mmdet3d_plugin/datasets/custom_nuscenes_dataset.py", line 114, in get_data_info
    prev_idx=info['prev'],
KeyError: 'prev'

The dataset is nuscenes-mini and I generated the pkl from mmdet3d rep. I wonder which version of pkl do you use? The same as in Cylinder3D or others like BEVFormer. Really appreciate your help。

Hi, thanks for your project.

I downloaded nuScenes-C dataset, but it contains only 6,019 images for each camera.

I want to generate fog, snow dataset for all nuScenes training data using your code.

Could you tell me which code I have to use?

Hi there,
Thank you all for your benchmarking work.
Just a quick question, I saw there is a nuscenes-c dataset available here: https://opendatalab.com/OpenDataLab/nuScenes-C/cli/main.
I am wondering if the corrupted dataset was generated from Nuscenes' original dataset or the densified one from SurroundOcc. (I saw there are two occupancy prediction baselines in this repo. However, their ground truth labels in their original paper are different.)
Thank you for your time.
Regards,
Jo

Welcome update to OpenMMLab 2.0

I am Vansin, the technical operator of OpenMMLab. In September of last year, we announced the release of OpenMMLab 2.0 at the World Artificial Intelligence Conference in Shanghai. We invite you to upgrade your algorithm library to OpenMMLab 2.0 using MMEngine, which can be used for both research and commercial purposes. If you have any questions, please feel free to join us on the OpenMMLab Discord at https://discord.gg/A9dCpjHPfE or add me on WeChat (ID: van-sin) and I will invite you to the OpenMMLab WeChat group.

Here are the OpenMMLab 2.0 repos branches:

Attention: please create a new virtual environment for OpenMMLab 2.0.

Following configs are the same. How did you test non-temporal model?
Temporal config:

zoo/BEVFormer/projects/configs/bevformer/bevformer_base.py

Non-temporal configs:

zoo/BEVFormer/projects/configs/bevformer/bevformer_base_no_temp.py