search

jzhzhang / 3DAwareNav

[CVPR 2023] We propose a framework for the challenging 3D-aware ObjectNav based on two straightforward sub-policies. The two sub-polices, namely corner-guided exploration policy and category-aware identification policy, simultaneously perform by utilizing online fused 3D points as observation.
MIT License
57 stars 3 forks source link

Environmental issues when I run the code #2

Closed peakfly closed 1 year ago

peakfly commented 1 year ago

Hello，Thanks for your work.I had some problems when I run the code What's the version of pytorch, torchvision,cudatoolkit and detectron2? Looking forward to your reply.

jzhzhang commented 1 year ago

You can follow the instructions of Habitat-sim(v0.2.2) and habitat-lab (v0.2.2) to finish the installation, which will also include install torch-related lib. And then you can install detectron2 based on your torch and cuda version by following detectron2 installation guide. Another option is to directly install detectorn2 from its source code.

peakfly commented 1 year ago
Thank you! I have followed the instructions of Habitat-sim (v0.2.2) and habitat-lab (v0.2.2).I have also installed detectorn2 from its source code by following git clone https://github.com/facebookresearch/detectron2.git python -m pip install -e detectron. The conda list is as follows： Name Version
python 3.7.16
torch 1.13.1
nvidia-cuda-runtime-cu11 11.7.99
detectron2 0.6
habitat 0.2.2
habitat-sim 0.2.2

But when I run the sh sh_train_mp3d.sh,the error occured: image The chatgpt said it was a problem with the python version.The python version should >=3.8. I have no idea.

peakfly commented 1 year ago

Could you give me some advice on the version of pytorch,cuda and python ? Thank you!