peiyunh/ff

Safe Local Motion Planning with Self-Supervised Freespace Forecasting, CVPR 2021

ff's Issues

Fail to compile code for Lidar point cloud ground segmentation

Thanks for sharing the code! When I run preprocess.py, I have the following error:

Traceback (most recent call last):
File "preprocess.py", line 11, in
from lib.grndseg import segmentation
ImportError: cannot import name 'segmentation' from 'lib.grndseg' (unknown location)

Does it mean I didn't compile the code in lib/grndseg successfully? Thanks!

Fail while running precast.py file

Hi, I'm a newbie studying about using deep learning to predict future trajectory on cars.
I have been studying your paper, and I am now starting to run your code.
I made a dockerfile to run this code by using nuscenes' mini data, and I succeed on running "preprocess.py",
but I failed while running "precast.py".

I got this error, " keyerror : 'train on all sweeps' ", and I would like to get some advice on how to solve this error.
Also if possible, could you explain the order in which code should be executed?

It would be a great honor to run your code by your help.

(base) root@b1d627403c9b:/usr/src/app# python -W ignore precast.py
Using /root/.cache/torch_extensions as PyTorch extensions root...
Detected CUDA files, patching ldflags
Emitting ninja build file /root/.cache/torch_extensions/raycaster/build.ninja...
Building extension module raycaster...
Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N)
[1/2] /usr/local/cuda/bin/nvcc -DTORCH_EXTENSION_NAME=raycaster -DTORCH_API_INCLUDE_EXTENSION_H -isystem /opt/conda/lib/python3.6/site-packages/torch/include -isystem /opt/conda/lib/python3.6/site-packages/torch/include/torch/csrc/api/include -isystem /opt/conda/lib/python3.6/site-packages/torch/include/TH -isystem /opt/conda/lib/python3.6/site-packages/torch/include/THC -isystem /usr/local/cuda/include -isystem /opt/conda/include/python3.6m -D_GLIBCXX_USE_CXX11_ABI=1 -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr -gencode=arch=compute_70,code=sm_70 -gencode=arch=compute_52,code=sm_52 -gencode=arch=compute_61,code=sm_61 -gencode=arch=compute_86,code=sm_86 -gencode=arch=compute_86,code=compute_86 -gencode=arch=compute_80,code=sm_80 -gencode=arch=compute_60,code=sm_60 -gencode=arch=compute_75,code=sm_75 --compiler-options '-fPIC' -std=c++14 -c /usr/src/app/lib/raycast/raycaster.cu -o raycaster.cuda.o
[2/2] c++ raycaster.o raycaster.cuda.o -shared -L/opt/conda/lib/python3.6/site-packages/torch/lib -lc10 -lc10_cuda -ltorch_cpu -ltorch_cuda -ltorch -ltorch_python -L/usr/local/cuda/lib64 -lcudart -o raycaster.so
Loading extension module raycaster...

Loading NuScenes tables for version v1.0-mini...
Loading nuScenes-lidarseg...
32 category,
8 attribute,
4 visibility,
911 instance,
12 sensor,
120 calibrated_sensor,
31206 ego_pose,
8 log,
10 scene,
404 sample,
31206 sample_data,
18538 sample_annotation,
4 map,
404 lidarseg,
Done loading in 0.375 seconds.

Reverse indexing ...
Done reverse indexing in 0.1 seconds.

Traceback (most recent call last):
File "precast.py", line 29, in
dataset = nuScenesDataset(nusc, "train", dataset_kwargs)
File "/usr/src/app/data.py", line 106, in init
self.train_on_all_sweeps = kwargs["train_on_all_sweeps"]
KeyError: 'train_on_all_sweeps'

Sample question

Thanks for sharing your code!
I see that your sampler.py code contains TODO statements. Does it mean that this code is not the final version?
I also visualized the results of the sampler code, and I can see that there is still a large gap between the clothoid curve and the straight line that has not been sampled.
How to solve this question?
thanks~~

ImportError: dynamic module does not define module export function (PyInit_segmentation)

Thanks for sharing your code!
I encounter the problem about compile grndseg

ImportError: dynamic module does not define module export function (PyInit_segmentation)

How to solve it? Thanks

where?when?what?

How do you deal with the final model evaluation

Hello there,
Thanks for your nice code.
I just found out that nuscenes didn't open source their test datasets' annotation.( for cheating avoiding)
And I checked the evaAI platform but didn't find evaluation possibility for planning.

How do you deal with this issue?
What do you use to finally evaluate your model?

Best regards,
RL