Troubleshooting¶
Common issues and solutions for NavArena installation, data generation, and evaluation.
Installation¶
gsplat Compilation Failed¶
If gsplat fails to compile, check:
- CUDA version matches PyTorch:
python -c "import torch; print(torch.version.cuda)" - C++ compiler is installed:
gcc --version - Sufficient disk space (compilation requires temporary space)
- First
import gsplattriggers JIT compilation which may take several minutes
CUDA Version Mismatch¶
Ensure the installed PyTorch version is compatible with your system CUDA driver:
nvidia-smi # Check max CUDA version supported by driver
python -c "import torch; print(torch.cuda.is_available()); print(torch.version.cuda)"
The driver's CUDA version must be ≥ the CUDA version PyTorch was compiled with.
open3d Installation Failed¶
open3d only supports Python 3.8-3.11, and some systems may lack dependencies:
Node.js / npm install Failed¶
The Web Viewer frontend requires Node.js. Install via conda:
If npm install fails, try clearing the cache:
conda env create fails due to network issues¶
A common cause is GitHub-based dependencies (like CLIP) failing to download. requirements-lock.txt has separated CLIP into an optional post-install step.
For China users, try a pip mirror:
pip install -i https://pypi.tuna.tsinghua.edu.cn/simple -r requirements-lock.txt
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
If CLIP installation fails, retry or use a proxy. CLIP is only needed for ObjectNav; PointNav/ImageNav/VLN work without it.
Permission Errors¶
Data Generation¶
Config or scene path not found¶
- Ensure
NAVARENA_DATA_DIRis set:echo $NAVARENA_DATA_DIR - Check that
scene_pathin your config matches the actual directory under$NAVARENA_DATA_DIR/assets/ - The config
scene_pathformat is{group}/{scene_id}(e.g.x2robot/17dc3367, where group is the scene group name)
No episodes generated / empty output¶
- Verify assets have completed preprocessing (manifest.json, aligned.ply, nav_map.pgm exist)
- Check that the scene has enough navigable area; some scenes may have too few valid spawn points
- Reduce
num_episodesfor a quick test (e.g.--num-episodes 5)
Evaluation¶
Evaluation returns empty results¶
- Ensure
eval_results_dirin config exists and is writable - Check that the dataset path in config points to valid Parquet data
- Verify
save_trajectories: truein evaluator config if you need replay videos
Agent connection failed (RemoteAgent / ViNT)¶
- For RemoteAgent: Ensure the remote service is running and the URL in config is correct
- For ViNT: Verify visualnav-transformer is cloned and diffusion_policy is installed; check that model weights are in
$NAVARENA_DATA_DIR/shared/visualnav-transformer/