KeyError: "NuScenesDataset: 'infos'"

[lijain]

I WANT TO PREDICETION : python tools/test.py configs/fcos3d/fcos3d_r101_caffe_fpn_gn-head_dcn_2x8_1x_nus-mono3d_finetune.py model/weights/fcos3d_r101_caffe_fpn_gn-head_dcn_2x8_1x_nus-mono3d_finetune_20210427_091419-35aaaad0.pth --show-dir /home/mnt_abc004-data/mq/Project/mmdetection3d-master/output/ --eval mAP

always no key "NuScenesDataset: 'infos'"，why???
my set

[Wuziyi616]

MMDet3D requires generating .pkl info files to provide the input path and annotations. Please follow the doc to prepare nuScenes data.

If you want to do inference on one image, see our demo using FCOS3D.

[lijain]

do not??

[Wuziyi616]

Can you post the entire error output message here? I am a bit confused about your question now.

[lijain]

ok，Yes, following the first question, nothing has changed in my order.The input test data format is coco. JSON.Always no key "NuscenesDataset: 'infos'" in test？

[Wuziyi616]

I think I got your point. The reason is that, the data under demo/data/ is for our demo program (e.g. FCOS3D). So they are slightly different from our generated info files used by test.py. You can't use it as input to test.py. If you only want to do prediction on that one image, try the demo file. If you want to test FCOS3D on the entire nuScenes dataset, you need to use create_data.py.

[lijain]

i get it.

python demo/mono_det_demo.py /home/mnt_abc004-data/mq/Project/mmdetection3d-master/demo/data/nuscenes/images/n015-2018-07-24-11-22-45+0800__CAM_BACK__1532402927637525.jpg /home/mnt_abc004-data/mq/Project/mmdetection3d-master/demo/data/nuscenes/n015-2018-07-24-11-22-45+0800__CAM_BACK__1532402927637525_mono3d.coco.json configs/fcos3d/fcos3d_r101_caffe_fpn_gn-head_dcn_2x8_1x_nus-mono3d_finetune.py model/weights/fcos3d_r101_caffe_fpn_gn-head_dcn_2x8_1x_nus-mono3d_finetune_20210427_091419-35aaaad0.pth --out-dir /home/mnt_abc004-data/mq/Project/mmdetection3d-master/output/
it have a problem:

looked for your issues, and the online search failed to solve.Please give directions, thank

[Wuziyi616]

That's so strange, I test the demo and it works smoothly. When do you download our repo? We have a update recently which may causes your bug. Maybe you can download the newest master branch and have a try?

[lijain]

I have one more question,
i test :python demo/pcd_demo.py demo/data/kitti/kitti_000008.bin configs/second/hv_second_secfpn_6x8_80e_kitti-3d-car.py data/predtrain_models/hv_second_secfpn_6x8_80e_kitti-3d-car_20200620_230238-393f000c.pth --out-dir data/output_result/
so result:

I keep getting errors when I visualize:python tools/misc/visualize_results.py ${CONFIG_FILE} --result ${RESULTS_PATH} --show-dir
${SHOW_DIR}
always error not .pkl file, Am I using the wrong command?
other You're using that version

[Wuziyi616]

Oh no no! tools/misc/visualize_results.py is used to visualize results saved by tools/test.py. The saved results from demo can be easily visualized using Meshlab, which is a commonly used 3D visualization software. You can simply download it from Internet (I think we already clarify this point in the doc).

[lijain]

Version 0.14.0 should not ask! Which version do you use?

[Wuziyi616]

MMDetection3D: 0.15.0 with MMDetection 2.14.0, MMSegmentation 0.14.1 and MMCV 1.3.9. Yes you need to update MMDetection3D to 0.15.0!

[lijain]

ok thank you.Yesterday I downloaded the master

[lijain]

I recompiled it to the environment you said. Run :python demo/mono_det_demo.py demo/data/nuscenes/n015-2018-07-24-11-22-45+0800__CAM_BACK__1532402927637525.jpg demo/data/nuscenes/n015-2018-07-24-11-22-45+0800__CAM_BACK__1532402927637525_mono3d.coco.json configs/fcos3d/fcos3d_r101_caffe_fpn_gn-head_dcn_2x8_1x_nus-mono3d_finetune.py weights/pretain/fcos3d_r101_caffe_fpn_gn-head_dcn_2x8_1x_nus-mono3d_finetune_20210427_091419-35aaaad0.pth --out-dir /home/mnt_abc004-data/mq/Project/mmdetection3d-15/result/output

There is a problem:

Searched the reason on the net, say :

Do I want to change the source code??

[Wuziyi616]

Sorry I didn't see this. Can you run python mmdet3d/utils/collect_env.py to show me your environment? Also what opencv/numpy/torch version are you using?

[lijain]

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Tue_Jun_12_23:07:04_CDT_2018
Cuda compilation tools, release 9.2, V9.2.148
Package Version Location

---

absl-py 0.13.0
addict 2.4.0
anyio 3.3.0
appdirs 1.4.4
argon2-cffi 20.1.0
astor 0.8.1
async-generator 1.10
asynctest 0.13.0
attrs 21.2.0
Babel 2.9.1
backcall 0.2.0
black 21.7b0
bleach 3.3.1
cached-property 1.5.2
cachetools 4.2.2
certifi 2021.5.30
cffi 1.14.6
charset-normalizer 2.0.3
click 8.0.1
codecov 2.1.11
colorama 0.4.4
coverage 5.5
cycler 0.10.0
Cython 0.29.24
dataclasses 0.6
debugpy 1.4.0
decorator 5.0.9
defusedxml 0.7.1
deprecation 2.1.0
descartes 1.1.0
entrypoints 0.3
fire 0.4.0
flake8 3.9.2
future 0.18.2
gast 0.2.2
google-auth 1.33.1
google-auth-oauthlib 0.4.4
google-pasta 0.2.0
grpcio 1.39.0
h5py 3.3.0
idna 3.2
imageio 2.9.0
importlib-metadata 3.10.1
iniconfig 1.1.1
interrogate 1.4.0
ipykernel 6.0.3
ipython 7.25.0
ipython-genutils 0.2.0
ipywidgets 8.0.0a5
isort 5.9.2
jedi 0.18.0
Jinja2 3.0.1
joblib 1.0.1
json5 0.9.6
jsonschema 3.2.0
jupyter 1.0.0
jupyter-client 6.1.12
jupyter-console 6.4.0
jupyter-core 4.7.1
jupyter-packaging 0.10.4
jupyter-server 1.10.0
jupyterlab 3.0.16
jupyterlab-pygments 0.1.2
jupyterlab-server 2.6.1
jupyterlab-widgets 2.0.0a2
Keras-Applications 1.0.8
Keras-Preprocessing 1.1.2
kiwisolver 1.3.1
kwarray 0.5.19
llvmlite 0.31.0
lyft-dataset-sdk 0.0.8
Markdown 3.3.4
MarkupSafe 2.0.1
matplotlib 3.4.2
matplotlib-inline 0.1.2
mccabe 0.6.1
mistune 0.8.4
mkl-fft 1.3.0
mkl-random 1.2.2
mkl-service 2.4.0
mmcv-full 1.3.9
mmdet 2.14.0
mmdet3d 0.15.0 /home/mnt_abc004-data/mq/Project/mmdetection3d-15
mmsegmentation 0.15.0
mypy-extensions 0.4.3
nbclassic 0.3.1
nbclient 0.5.3
nbconvert 6.1.0
nbformat 5.1.3
nest-asyncio 1.5.1
networkx 2.2
notebook 6.4.0
numba 0.48.0
numpy 1.19.5
nuscenes-devkit 1.1.5
oauthlib 3.1.1
olefile 0.46
open3d 0.13.0
opencv-python 4.2.0.32
opt-einsum 3.3.0
ordered-set 4.0.2
packaging 21.0
pandas 1.3.0
pandocfilters 1.4.3
parso 0.8.2
pathspec 0.9.0
pexpect 4.8.0
pickleshare 0.7.5
Pillow 8.3.1
pip 21.1.3
plotly 5.1.0
pluggy 1.0.0.dev0
plyfile 0.7.4
prettytable 2.1.0
prometheus-client 0.11.0
prompt-toolkit 3.0.19
protobuf 4.0.0rc2
ptyprocess 0.7.0
py 1.10.0
pyasn1 0.4.8
pyasn1-modules 0.2.8
pycocotools 2.0.1
pycodestyle 2.7.0
pycparser 2.20
pyflakes 2.3.1
Pygments 2.9.0
pyparsing 2.4.7
pyquaternion 0.9.9
pyrsistent 0.18.0
pytest 6.2.4
pytest-cov 2.12.1
pytest-runner 5.3.1
python-dateutil 2.8.2
pytz 2021.1
PyWavelets 1.1.1
PyYAML 5.4.1
pyzmq 22.1.0
qtconsole 5.1.1
QtPy 1.9.0
regex 2021.7.6
requests 2.26.0
requests-oauthlib 1.3.0
requests-unixsocket 0.2.0
rsa 4.7.2
scikit-image 0.18.2
scikit-learn 0.24.2
scipy 1.4.1
Send2Trash 1.7.1
setuptools 52.0.0.post20210125
Shapely 1.8a2
six 1.16.0
sniffio 1.2.0
tabulate 0.8.9
tenacity 8.0.1
tensorboard 2.1.1
tensorboard-data-server 0.6.1
tensorboard-plugin-wit 1.8.0
tensorflow-estimator 2.1.0
tensorflow-gpu 2.1.0
termcolor 1.1.0
terminado 0.10.1
terminaltables 3.1.0
testpath 0.5.0
threadpoolctl 2.2.0
tifffile 2021.7.2
toml 0.10.2
tomli 1.0.4
tomlkit 0.7.2
torch 1.7.0
torchaudio 0.7.0a0+ac17b64
torchvision 0.8.0
tornado 6.1
tqdm 4.61.2
traitlets 5.0.5
trimesh 2.35.39
typed-ast 1.4.3
typing-extensions 3.10.0.0
ubelt 0.9.5
urllib3 1.26.6
waymo-open-dataset-tf-2-1-0 1.2.0
wcwidth 0.2.5
webencodings 0.5.1
websocket-client 1.1.0
Werkzeug 2.0.1
wheel 0.36.2
widgetsnbextension 4.0.0a3
wrapt 1.12.1
xdoctest 0.15.5
yapf 0.31.0
zipp 3.5.0
Is my CUDA version low?Or should I compile from source code for MMDetection etc

[Wuziyi616]

Hi @lijain, I really need your output of this command python mmdet3d/utils/collect_env.py. Please also post it here, because it will show something like GCC version, etc.

[lijain]

I'm sorry. It's a bit of a headache

[lijain]

I recompiled the environment, but I still have this problem:

Should I change mmDetection3D version, please? Thanks for your guidance.I hope you have time to guide me

[Wuziyi616]

This is really weird. The bug is from cv2 not our openmmlab packages. If you enter a command line of python and import cv2 and PyTorch at the same time, do you have such bug? (Try both the import order) I don't think this is a bug related to our mmdet3d.

[lijain]

I also find it strange that my previous installations with MMDetection have been very easy to install.I felt a little bit difficult this time. After checking the problem, it said that Opencv-Python conflicted with Python version. This problem also appeared when I used a single votenet project before.But I worked it out, and this time it didn't work either way. It was awkward.
I will try mmDection3D 0.14.0 tomorrow, it is almost the third version, master, 0.15.0 are not finished

[Wuziyi616]

Yeah I agree, maybe this is a opencv and mmdet problem.

[Wuziyi616]

For example, since you have installed mmdet, can you try import mmdet and test their demo? If that also fails, then it's not our mmdet3d. Instead, you should raise an issue in their repo.

[lijain]

I tested the environment:

[Wuziyi616]

What about import cv2 and import torch?

[lijain]

I recompiled it, using the original MMCV. Run python mmdet3d/utils/collect_env. Py

is ok ??
but run :python demo/mono_det_demo.py demo/data/nuscenes/n015-2018-07-24-11-22-45+0800__CAM_BACK__1532402927637525.jpg demo/data/nuscenes/n015-2018-07-24-11-22-45+0800__CAM_BACK__1532402927637525_mono3d.coco.json configs/fcos3d/fcos3d_r101_caffe_fpn_gn-head_dcn_2x8_1x_nus-mono3d_finetune.py weights/pretain/fcos3d_r101_caffe_fpn_gn-head_dcn_2x8_1x_nus-mono3d_finetune_20210427_091419-35aaaad0.pth --out-dir /home/mnt_abc004-data/mq/Project/mmdetection3d-15/result/output

[Wuziyi616]

I got the problem! You can see that MMCV CUDA compiler is 10.2, which is incompatible with your 9.2. Maybe you can download a pre-compiled MMCV or compile it with 9.2?

e.g. pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu92/torch1.6.0/index.html

If that doesn't work, you should compile mmcv with your cuda 9.2.

[lijain]

I changed, or appear the above problem

[Wuziyi616]

I really think this is an opencv-related problem, not ours. One last try, can you create a clean conda environment of python 3.7.9, with torch 1.5.0, mmcv 1.3.9, mmdet 2.14.0, mmseg 0.14.1 and mmdet3d 0.15.0 and test if it works? If that fails too, then I am afraid I can't work it out either... Really sorry about that.

[lijain]

What's your CUDA? Can you cut a picture of the conda environment installation package? I'll install it as you can run it

[lijain]

First of all, thank you Wuziyi616 for your patience. Finally, after many compilations, it was found that the main reason was that there were many installed libraries and some libraries had version conflicts (such as Numpy, Scikit-Learn, etc.). Share my own library version here, CUDA ==9.2
Package Version Location

---

absl-py 0.13.0
addict 2.4.0
anyio 3.3.0
appdirs 1.4.4
argon2-cffi 20.1.0
astor 0.8.1
async-generator 1.10
asynctest 0.13.0
attrs 21.2.0
Babel 2.9.1
backcall 0.2.0
black 21.7b0
bleach 3.3.1
cached-property 1.5.2
cachetools 4.2.2
certifi 2021.5.30
cffi 1.14.6
charset-normalizer 2.0.3
click 8.0.1
codecov 2.1.11
colorama 0.4.4
coverage 5.5
cycler 0.10.0
Cython 0.29.24
debugpy 1.4.1
decorator 5.0.9
defusedxml 0.7.1
deprecation 2.1.0
descartes 1.1.0
entrypoints 0.3
fire 0.4.0
flake8 3.9.2
gast 0.2.2
google-auth 1.33.1
google-auth-oauthlib 0.4.4
google-pasta 0.2.0
grpcio 1.39.0
h5py 3.3.0
idna 3.2
imageio 2.9.0
importlib-metadata 3.10.1
iniconfig 1.1.1
interrogate 1.4.0
ipykernel 6.0.3
ipython 7.25.0
ipython-genutils 0.2.0
ipywidgets 7.6.3
isort 5.9.2
jedi 0.18.0
Jinja2 3.0.1
joblib 1.0.1
json5 0.9.6
jsonschema 3.2.0
jupyter 1.0.0
jupyter-client 6.1.12
jupyter-console 6.4.0
jupyter-core 4.7.1
jupyter-packaging 0.10.4
jupyter-server 1.10.1
jupyterlab 3.0.16
jupyterlab-pygments 0.1.2
jupyterlab-server 2.6.1
jupyterlab-widgets 1.0.0
Keras-Applications 1.0.8
Keras-Preprocessing 1.1.2
kiwisolver 1.3.1
kwarray 0.5.19
llvmlite 0.31.0
lyft-dataset-sdk 0.0.8
Markdown 3.3.4
MarkupSafe 2.0.1
matplotlib 3.4.2
matplotlib-inline 0.1.2
mccabe 0.6.1
mistune 0.8.4
mkl-fft 1.3.0
mkl-random 1.2.2
mkl-service 2.4.0
mmcv-full 1.3.9 /home/mnt_abc004-data/mq/Project/mmdetection3d-15/mmcv
mmdet 2.14.0
mmdet3d 0.15.0 /home/mnt_abc004-data/mq/Project/mmdetection3d-15
mmsegmentation 0.14.1
mypy-extensions 0.4.3
nbclassic 0.3.1
nbclient 0.5.3
nbconvert 6.1.0
nbformat 5.1.3
nest-asyncio 1.5.1
networkx 2.2
notebook 6.4.0
numba 0.48.0
numpy 1.20.3
nuscenes-devkit 1.1.5
oauthlib 3.1.1
olefile 0.46
open3d 0.13.0
opencv-python 4.5.3.56
opt-einsum 3.3.0
ordered-set 4.0.2
packaging 21.0
pandas 1.3.1
pandocfilters 1.4.3
parso 0.8.2
pathspec 0.9.0
pexpect 4.8.0
pickleshare 0.7.5
Pillow 8.3.1
pip 21.1.3
plotly 5.1.0
pluggy 0.13.1
plyfile 0.7.4
prettytable 2.1.0
prometheus-client 0.11.0
prompt-toolkit 3.0.19
protobuf 3.17.3
ptyprocess 0.7.0
py 1.10.0
pyasn1 0.4.8
pyasn1-modules 0.2.8
pycocotools 2.0.2
pycodestyle 2.7.0
pycparser 2.20
pyflakes 2.3.1
Pygments 2.9.0
pyparsing 2.4.7
pyquaternion 0.9.9
pyrsistent 0.18.0
pytest 6.2.4
pytest-cov 2.12.1
pytest-runner 5.3.1
python-dateutil 2.8.2
pytz 2021.1
PyWavelets 1.1.1
PyYAML 5.4.1
pyzmq 22.1.0
qtconsole 5.1.1
QtPy 1.9.0
regex 2021.7.6
requests 2.26.0
requests-oauthlib 1.3.0
requests-unixsocket 0.2.0
rsa 4.7.2
scikit-image 0.18.2
scikit-learn 0.24.2
scipy 1.6.2
Send2Trash 1.7.1
setuptools 52.0.0.post20210125
Shapely 1.7.1
six 1.16.0
sniffio 1.2.0
tabulate 0.8.9
tenacity 8.0.1
tensorboard 2.1.1
tensorflow-estimator 2.1.0
tensorflow-gpu 2.1.0
termcolor 1.1.0
terminado 0.10.1
terminaltables 3.1.0
testpath 0.5.0
threadpoolctl 2.2.0
tifffile 2021.7.2
toml 0.10.2
tomli 1.1.0
tomlkit 0.7.2
torch 1.5.0
torchvision 0.6.0a0+82fd1c8
tornado 6.1
tqdm 4.61.2
traitlets 5.0.5
trimesh 2.35.39
typed-ast 1.4.3
typing-extensions 3.10.0.0
ubelt 0.9.5
urllib3 1.26.6
waymo-open-dataset-tf-2-1-0 1.2.0
wcwidth 0.2.5
webencodings 0.5.1
websocket-client 1.1.0
Werkzeug 2.0.1
wheel 0.36.2
widgetsnbextension 3.5.1
wrapt 1.12.1
xdoctest 0.15.5
yapf 0.31.0
zipp 3.5.0

[Wuziyi616]

@lijain So you can successfully run our demo now?

[lijain]

yes ,thanks you

[Wuziyi616]

Glad it helps!

[FlyingAnt2018]

yes ,thanks you

hello, where to find *.coco.json files for different images and different dataset ？